NCAT Report 11-01 LABORATORY REFINEMENT AND FIELD VALIDATION OF 4.75 mm SUPERPAVE DESIGNED ASPHALT MIXTURES Volume I: Final Report

By: Randy C. West Michael A. Heitzman David M. Rausch Grant Julian March 2011

i

LABORATORY REFINEMENT AND FIELD VALIDATION OF 4.75 mm SUPERPAVE DESIGNED ASPHALT MIXTURES

Volume I: Final Report

by

Randy C. West Director

Michael A. Heitzman

Assistant Director

Grant Julian Assistant Research Engineer

National Center for Asphalt Technology

Auburn, Alabama

and

Michael Rausch Skanska Infrastructure Development AB

(former graduate student)

March 2011

ii

DISCLAIMER

The contents of this report reflect the views of the authors who are solely responsible for the facts and accuracy of the data presented herein. The contents do not necessarily reflect the official view and policies of the National Center for Asphalt Technology at Auburn University. This report does not constitute a standard, specification, or regulation.

ACKNOWLEDGMENTS

The authors would like to thank the Alabama Department of Transportation, the Connecticut Department of Transportation, the Florida Department of Transportation, the Minnesota Department of Transportation, the Missouri Department of Transportation, the New Hampshire Department of Transportation, the Tennessee Department of Transportation, the Virginia Department of Transportation, and the Wisconsin Department of Transportation for their support and cooperation with this research. Special thanks are given to Dr. Osamu Takahashi of Nagaoka University of Technology for all of his efforts in the laboratory research for this project.

iii

LIST OF ACRONYMS/ABBREVIATIONS

APA – Asphalt Pavement Analyzer

AASHTO – American Association of State Highway and Transportation Officials

ASTM – Association of Standards for Testing and Materials

CTM – Circular Texture Meter

D:B ratio – Dust-to-Binder Ratio

DFT – Dynamic Friction Tester

ESAL – Equivalent Single-Axle Loads

FAA – Fine Aggregate Angularity

FE– Fracture Energy

FM – Fineness Modulus

G*- Complex Shear Modulus

Gmm – Mix Theoretical Maximum Specific Gravity

Gmb – Mix Bulk Specific Gravity

HMA – Hot-Mix Asphalt

IDT- Indirect Tension

LVDTs – Linear Variable Differential Transducers

Manf-sand – Manufactured Sand

MPD– Mean Profile Depth

MVP – Mixture Verification Tester

NCAT – National Center for Asphalt Technology

Ndes – Number of Gyrations at Design Level

Nini – Number of Gyrations at Initial Compaction

NMAS – Nominal Maximum Aggregate Size

P-075 – Percent Passing the 0.075 mm Sieve

Pb – Percent Binder (preferable over “AC”)

Pbe – Percent of Effective Binder

PG – Performance Grade (pertaining to asphalt binder)

QC/QA – Quality Control/Quality Assurance

iv

RAP – Reclaimed Asphalt Pavement

SE – Sand Equivalent

SGC – Superpave Gyrator Compactor

TSR – Tensile Strength Ratio

Va – Air Voids

Vbe – Volume of Effective Binder

VFA – Voids Filled with Asphalt

VMA – Voids in Mineral Aggregate

WMA – Warm-Mix Asphalt

WSD – Washed Sand

Statistical Acronyms

ADJ SS – Adjusted Sum of Squares

ADJ MS – Adjusted Mean Square

ANOVA – Analysis of Variance

P – p-value

S – Standard Deviations

SE – Squared Error

SS – Sum of Squares

SS sr – Reduced Sum of Squares

X - Mean

v

TABLE OF CONTENTS

PAGE CHAPTER 1 INTRODUCTION 1 1.1 Objective 2 1.2 Scope 2 CHAPTER 2 BACKGROUND 3 2.1 Development of Mix Design Criteria 3 2.2 Use of Screenings to Produce HMA Mixtures 4 2.3 Low Volume Road Applications 6 2.4 Leveling and Patching 7 2.5 Thin Overlays and Surface Mixtures 8 2.6 NCAT Survey 9 CHAPTER 3 RESEARCH PLAN 13 3.1 Test Methods 14 3.1.1 Aggregate Tests 14 3.1.2 Mix Designs 15 3.1.3 Performance Tests 16

3.1.3.1 Permanent Deformation 16 3.1.3.2 Moisture Damage Potential 18 3.1.3.3 Permeability 18 3.1.3.4 Cracking Resistance 19 CHAPTER 4 RESULTS AND ANALYSIS 23 4.1 Mix Design Results 25 4.1.1 Optimum Asphalt Content 25 4.1.2 VMA 31 4.1.3 VFA 34 4.1.4 Percent of Gmm at Ninitial 37 4.1.5 Dust-to-Binder Ratio and Film Thickness 39 4.1.6 Aggregate Properties (Gradation, SE, FAA) 40 4.2 Performance Tests 45 4.2.1 MVT Rut Depths 45 4.2.2 Tensile Strength Ratio 51 4.2.3 Fracture Energy Density Ratio 56 4.2.4 Permeability 61 4.3 Baseline Mixtures 64 4.4 Review of AASHTO Specifications 66 4.4.1 AASHTO Gradation Limits 67 4.4.2 Sand Equivalent 68 4.4.3 Dust-to-Binder Ratio 68 4.4.4 Fine Aggregate Angularity 69 4.4.5 Percent of Gmm at Nini 69

vi

4.4.6 Volumetric Requirements 70 4.4.7 Summary of Recommendations for Revising 4.75 mm NMAS Mix Design Criteria 71 CHAPTER 5 VALIDATION OF PROPOSED CRITERIA BY PLANT PRODUCTION AND CONSTRUCTION 73 5.1 Alabama Field Validation Project 74 5.1.1 Project Description 74 5.1.2 Mix Design 75 5.1.3 Sampling and Testing Summary 75 5.1.4 Analysis of Mix Design and Production Test Results 76 5.2 Missouri Field Validation Project 77 5.2.1 Project Description 77 5.2.2 Mix Design 78 5.2.3 Sampling and Testing Summary 78 5.2.4 Analysis of Mix Design and Production Test Results 79 5.3 Tennessee Field Validation Project 80 5.3.1 Project Description 80 5.3.2 Mix Design 81 5.3.3 Sampling and Testing Summary 81 5.3.4 Analysis of Mix Design and Production Test Results 83 5.3.4.1 Virgin Aggregate Mixture 83 5.3.4.2 Mixture with 15% RAP 84 5.3.4.3 Virgin Mix-RAP Mix Comparison 86 5.4 Minnesota Field Validation Project 86 5.4.1 Project Description 86 5.4.2 Mix Design 86 5.4.3 Sampling and Testing Summary 86 5.4.4 Analysis of Mix Design and Production Test Results 88 5.5 Summary Analysis of Field Validation Results 89 5.5.1 Compliance with Mix Design Standards 89 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS 95 6.1 Conclusions 95 6.2 Recommendations 96 6.3 Recommended Applications, Advantages and Disadvantages 7 REFERENCES 99 APPENDIX 101 *Volume II: Details of Experimental 4.75 mm Mix Designs is available upon request.

vii

LIST OF TABLES PAGE

TABLE 2.1 Gradations and Properties of Screenings 5 TABLE 2.2 Typical Gradations of Dense-Graded Patching Mixtures 8 TABLE 2.3 Summary of Responses for States Having a 4.75 mm-like Mix 10 TABLE 2.4 Approximate Production of 4.75 mm NMAS Mixtures 11 TABLE 3.1 Original Mix Test Matrix 14 TABLE 3.2 4.75 mm Superpave Control Points 15 TABLE 4.1 Mix Design Volumetric Properties 23 TABLE 4.2 Materials and Stockpile Percentages for Laboratory Mixtures 24 TABLE 4.3 Blend Gradation for Laboratory Mixtures 25 TABLE 4.4 Analysis of Variance for Effective Asphalt Content 27 TABLE 4.5 Mix Design Comparisons for Ndes=50 (4–6% Air Voids) 28 TABLE 4.6 Mix Design Comparison for 4% Air Voids (50–75 Gyrations) 29 TABLE 4.7 Mix Design Comparison for Ndes=75 (4–6% Air Voids) 29 TABLE 4.8 Analysis of Variance for VMA 33 TABLE 4.9 AASHTO Specifications for 4.75 mm NMAS Superpave Mixtures 35 TABLE 4.10 Analysis of Variance for VFA 35 TABLE 4.11 Descriptive Statistics for %Gmm @ Nini 37 TABLE 4.12 Analysis of Variance for %Gmm @ Nini 38 TABLE 4.13 Analysis of Variance for Dust-to-Binder Ratio 40 TABLE 4.14 Pearson Coefficients for Sand Equivalence 45 TABLE 4.15 Rut Depth and Mixture Properties for All Mix Designs 46 TABLE 4.16 Fracture Energy Results for Laboratory Mixtures 57 TABLE 4.17 Pearson Correlation Coefficients and p-values for Linear

Relationships with Fracture Energy Ratio 59 TABLE 4.18 Fracture Energy Density Data for Baseline Mixtures 61 TABLE 4.19 Permeability and Mixture Data for Laboratory Mixtures 62 TABLE 4.20 Mixture Properties and Performance Data for Baseline Mixtures 65 TABLE 4.21 AASHTO Mixture Criteria for 4.75 mm NMAS Superpave Asphalt Mixtures 67 TABLE 4.22 Proposed Design Criteria for 4.75 mm NMAS Superpave Mixtures 72 TABLE 5.1 Alabama Validation Project 4.75 mm Mix Design Summary 75 TABLE 5.2 NCAT Field Sampling and Testing for the Alabama Validation

Project 76 TABLE 5.3 Missouri Validation Project 4.75 mm Mix Design Summary 78 TABLE 5.4 NCAT Field Sampling and Testing for the Missouri Validation

Project 79 TABLE 5.5 Tennessee Validation Project 4.75 mm Virgin Mix Design

Summary 81 TABLE 5.6 Tennessee Validation Project 4.75 mm RAP Mix Design

Summary 82

viii

TABLE 5.7 NCAT Field Sampling and Testing for the Tennessee Validation Project (Virgin Mix) 82

TABLE 5.8 NCAT Field Sampling and Testing for the Tennessee Validation Project (15% RAP Mix) 83

TABLE 5.9 Minnesota Validation Project 4.75 mm Mix Design Summary 87 TABLE 5.10 NCAT Field Sampling and Testing for the Minnesota Validation

Project 87 TABLE 5.11 Summary of Mix Designs for Validation Projects 89 TABLE 5.12 Summary of Plant-Produced Mixes for Validation Projects 91 TABLE 5.13 Mix Design Criteria Validation Summary 92 TABLE 6.1 Proposed Design Criteria for 4.75 mm NMAS Superpave Design Mixtures 97

ix

LIST OF FIGURES PAGE FIGURE 2.1 Map of Respondents to NCAT Survey 9 FIGURE 3.1 Gradations for State Mixtures 15 FIGURE 3.2 Material Verification Tester 17 FIGURE 3.3 APA Rut Depths versus MVT Rut Depths 18 FIGURE 3.4 Best Fit Curves for In-Place Air Voids versus Permeability of Different NMAS 19 FIGURE 3.5 Area Under Stress-Strain Curve at Point of Fracture 20 FIGURE 3.6 Determination of Point of Fracture 21 FIGURE 3.7 Relationship Between Field Fatigue Performance and Fracture Energy 21 FIGURE 3.8 Instron Indirect Tension Tester 22 FIGURE 4.1 Optimum Asphalt Content 26 FIGURE 4.2 Effective Asphalt Content 27 FIGURE 4.3 Mean Effective Asphalt Content for 4% and 6% Air Voids (Ndes=50) 30 FIGURE 4.4 Mean Effective Asphalt Content for Ndes=50 and 75 (4% Air Voids) 30 FIGURE 4.5 Mean Effective Asphalt Content for 4% and 6% Air Voids (Ndes=75) 31 FIGURE 4.6 VMA Results for Each Mix Design 32 FIGURE 4.7 Mean VMA for 4% and 6% Air Voids (Ndes=50) 33 FIGURE 4.8 Mean VMA for 4% and 6% Air Voids (Ndes=75) 34 FIGURE 4.9 Mean VMA for Ndes=50 and 75 at 4% Design Air Voids 34 FIGURE 4.10 Mean VFA for 4% and 6% Air Voids (Ndes=50) 36 FIGURE 4.11 Mean VFA for Ndes=50 and 75 (4% Air Voids) 36 FIGURE 4.12 Mean VFA for 4% and 6% Air Voids (Ndes=75) 37 FIGURE 4.13 Mean %Gmm @ Nini for 4% and 6% Air Voids (Ndes=50) 38 FIGURE 4.14 Mean %Gmm @ Nini for Ndes=50 and 75 (4% Air Voids) 39 FIGURE 4.15 Mean %Gmm @ Nini for 4% and 6% Air Voids (Ndes=75 39 FIGURE 4.16 VMA Versus Percent Volume Effective Asphalt 41 FIGURE 4.17 Fineness Modulus Versus VMA for Over and Under 10% Passing

the 0.075 mm Sieve 41 FIGURE 4.18 VMA Versus Percent Passing 1.18 mm Sieve for Over and Under 10% Passing the 0.075 mm Sieve 42 FIGURE 4.19 FAA Versus VMA for Ndes= 50 and Design Air Voids = 4% 44 FIGURE 4.20 FAA Versus Gmm @ Nini for Ndes=50 and Design Air Voids = 4% 44 FIGURE 4.21 Mean Difference in VMA for Mixtures that Terminated Early and Completed 8000 Cycles on MVT 47 FIGURE 4.22 VMA Versus Cycles to Termination 48 FIGURE 4.23 VMA Versus Rutting Rate by Design Air Voids 48 FIGURE 4.24 Volume of Effective Asphalt Versus Rutting Rate by Design Air Voids 49 FIGURE 4.25 Vbe Versus Rutting Rate for All Mixtures 49

x

FIGURE 4.26 Vbe Versus Rutting Rate for All Mixtures Sorted by Percent Natural Sand 50 FIGURE 4.27 Vbe Versus Rutting Rate for All Mixtures Sorted by FAA 50 FIGURE 4.28 Relationship Between MVT Rut Depths at 120 lb, 120 psi to MVT Rut Depths at 100 lb, 100 psi 51 FIGURE 4.29 Tensile Strength Ratios for 29 Mix Designs 52 FIGURE 4.30 Tensile Strength for Conditioned and Unconditioned Samples 53 FIGURE 4.31 VMA Versus Tensile Strength Ratio 54 FIGURE 4.32 Effective Asphalt Content by Volume Versus Tensile Strength

Ratio 54 FIGURE 4.33 Relationship with Percent Natural Sand in Blended Aggregate and TSR for 50 Gyration 4% Air Void Mix Design 55 FIGURE 4.34 Dry Strength versus Film Thickness 55 FIGURE 4.35 Fracture Energy Ratio for Laboratory Mixtures 58 FIGURE 4.36 Fracture Energy Ratio Versus Vbe 58 FIGURE 4.37 Fracture Energy Ratio Versus Dust-to-Binder Ratio 59 FIGURE 4.38 Fracture Energy Ratio Versus VMA 59 FIGURE 4.39 Fracture Energy Ratio Versus VFA 60 FIGURE 4.40 Fracture Energy Ratio Versus Film Thickness 60 FIGURE 4.41 Permeability for Laboratory Mixtures 63 FIGURE 4.42 FAA Versus Permeability 64 FIGURE 4.43 Gradations for Baseline Mixtures 66 FIGURE 4.44 Vbe Versus Percent Passing the 1.18 mm Sieve for Over and Under 10% Passing the 0.075 mm Sieve 68 FIGURE 4.45 Rutting Rate Versus Dust-to-Binder Ratio 69 FIGURE 4.46 %Gmm @ Nini Versus Rutting Rate for 6% Design Air Voids 70 FIGURE 5.1 Average Plant-Production Gradation for Field Validation Projects 90 FIGURE 5.2 Average AASHTO T 283 Results for Plant-Produced Mixtures 93 FIGURE 5.3 In-Place Air Voids Versus Permeability for 4.75 mm Mixtures 94

1

LABORATORY REFINEMENT AND FIELD VALIDATION OF 4.75 mm SUPERPAVE DESIGNED ASPHALT MIXTURES

CHAPTER 1 INTRODUCTION 1.0 INTRODUCTION Until recently, 9.5 mm was the smallest nominal maximum aggregate size (NMAS) used in Superpave mix design criteria. In 2002, the National Center for Asphalt Technology (NCAT) completed a research study to develop Superpave mix design criteria for 4.75 mm NMAS mix (1). With the help of this research, the Superpave Mixture/Aggregate Expert Task Group recommended to AASHTO the addition of 4.75 mm NMAS mixes to the Superpave mix design system.

Many state agencies have expressed an interest in using 4.75 mm NMAS Superpave designed mixtures for thin lift applications, thin lift maintenance, and leveling courses to decrease construction time, to provide a use for screening stockpiles, and to provide an economical surface mix for low volume roads. A 2004 NCAT survey of state agencies indicated that several states were using 4.75 mm NMAS mixtures or mix types reasonably close to the AASHTO criteria for 4.75 mm mixes. The survey also confirmed that more agencies were interested in using 4.75 mm NMAS mixes in the future.

Although the original NCAT study on 4.75 mm mixes (1) provided initial criteria for 4.75 mm NMAS Superpave mixes, it was recommended that the mix design criteria be further refined in the laboratory and field validated. Criteria refinement was recommended in the following areas:

1. Minimum VMA criteria and dust-to-binder ratio (D:B ratio) requirements 2. Maximum VMA requirements 3. %Gmm @ Nini criteria 4. Aggregate properties 5. Binder contents and design air voids (Va) level (e.g., 4%) 6. Enhanced performance with the use of polymer modified binders

Since the original study (1) was performed with two aggregate sources, it was also

recommended that the refinement study incorporate materials from more states to obtain a large range of aggregate types.

In 2004, a pooled fund study was initiated to refine mix design criteria for 4.75 mm NMAS Superpave designed mixes and field validate design criteria. Nine states were participants in this study: Florida, Virginia, Missouri, Minnesota, Alabama, Tennessee, Wisconsin, New Hampshire, and Connecticut. Research began at NCAT in winter of 2005 for the laboratory refinement phase of this project.

2

1.1 Objective The main objective of this study is to refine the current procedures and criteria for 4.75 mm NMAS Superpave designed mixtures. Specifically, the criteria to be refined were

• Minimum Voids in the Mineral Aggregate (VMA) requirements and a workable range for Voids Filled with Asphalt (VFA)

• Percent of Mixture Theoretical Maximum Specific Gravity at Number of Gyrations at Initial Compaction (%Gmm @ Nini) requirements

• Aggregate properties such as Sand Equivalence (SE) and Fine Aggregate Angularity (FAA) of mixture

• Appropriate design air void content for a given compaction effort • D:B ratio requirements • A recommendation on the use of modified binders to enhance performance of 4.75 mm

NMAS asphalt mixtures 1.2 Scope A literature review was completed to understand the history and practical use of 4.75 mm NMAS Superpave designed mixtures. Next, a laboratory test plan was created. This test plan included performing numerous Superpave mix designs using materials provided by each state participating in the study. For each material and mix design, aggregate properties were measured, optimum asphalt content was determined for a given compaction effort and design Va, and performance tests were conducted to determine how well the mixtures performed for a given set of properties. The results of these mix designs were compared with the current AASHTO specification for 4.75 mm NMAS Superpave mixtures. These comparisons coupled with the results of the performance tests are used to measure the appropriateness of the current specification and to make improvement recommendations.

This report presents the findings of the pooled-fund 4.75 mm Superpave refinement project.

3

CHAPTER 2 BACKGROUND 2.0 BACKGROUND 2.1 Development of Mix Design Criteria for 4.75 mm Superpave Mixes In 2002, Cooley et al. (1), published research conducted at NCAT on the topic of specifications for 4.75 mm Superpave mixtures. The objective of this study was to develop mix design criteria for 4.75 mm NMAS mixtures. The research targeted the criteria of gradation controls and volumetric property requirements. Only two aggregate types were used in this study: granite and a limestone. Three gradations were evaluated for each gradation type: fine, medium and coarse. For each mixture, asphalt content was determined for 4% and 6% Va at a compactive effort of 75 gyrations. To analyze rutting susceptibility, an Asphalt Pavement Analyzer (APA) was used. This study confirmed that 4.75 mm NMAS can be successfully designed in the laboratory. Optimum binder contents of the designed mixes were affected by aggregate type, shape of the gradation curve, dust content, and designed Va. Voids in mineral aggregate values were affected by aggregate type, shape of the gradation curve, and dust content. APA rutting results were very high for the 4.75 mm mixtures as a result of the high binder contents required to meet the target Va of 4.0%. Although a good relationship was evident between VMA and D:B ratio, the finding could be limited by the two aggregate sources used. The results of the study, combined with existing mix design criteria from Maryland and Georgia, led to the following recommendations for designing 4.75 mm mixtures:

• Gradations for 4.75 mm NMAS mixes should be controlled on the 1.18 mm and 0.075 mm sieves.

• On the 1.18 mm sieve, the gradation control points are recommended as 30% to 54%. • On the 0.075 mm sieve, the control points are recommended as 6% to 12%. • A target designed Va of 4.0% should be used. • For all traffic levels, a minimum VMA of 16.0 should be utilized. • For mixes designed at 50 gyrations (very low traffic applications), no maximum VMA

criteria should be utilized. For mixes designed at 75 gyrations and above, a maximum VMA value of 18% is rational.

• For mixes designed at 75 gyrations and above, VFA criteria should be 75% to 78%. • Percent Gmm @ Nini values currently specified in AASHTO MP2-01 for the different

traffic levels are recommended. • Criteria for D:B ratio are recommended as 0.9 to 2.2.

Cooley provided the draft mix design criteria for 4.75 mm NMAS Superpave mixtures, but

recommended that mix design procedures be refined in the following areas:

1. Minimum VMA Criterion and D:B Ratio Requirements: Laboratory work is needed to evaluate the aging characteristics of 4.75 mm NMAS mixes. The minimum criterion of 16.0% was selected based upon Maryland and Georgia minimum binder contents and gradation specifications on similar mixes. An evaluation of the maximum D:B ratio requirement should be included in this work.

4

2. Maximum VMA Criteria: High optimum binder contents were identified as the primary cause of excessive laboratory rutting. For this reason, a maximum VMA criteria of 18.0% was recommended. This value needs to be validated in the laboratory by designing numerous mixes with a wide range of aggregate types to further evaluate the relationship between VMA and rut resistance.

3. %Gmm @ Nini Criteria: Within this study, two high quality aggregates were utilized. None of the 36 designed mixes failed the %Gmm @ Nini criteria for a 75 gyration design (90.5%). Additional work needs to be conducted that incorporates various percentages of natural, rounded sand to evaluate the applicability of %Gmm @ Nini requirements within the mix design system.

4. Aggregate Properties: Both of the aggregates used in this study had FAA values in excess of 45%. Refining the desired FAA values for different design levels should be further evaluated. Research is also needed to quantify an acceptable aggregate toughness and resistance to abrasion.

5. Air Void Criteria: To avoid excessive binder contents, field work should verify if 4.75 mm NMAS mixes can be designed at a single Va level (e.g., 4%) and result in satisfactory performance or determine if a design Va range criteria are needed.

6. Use of Polymer Modified Binders: Within a refinement study, some polymer modified binders should be included to evaluate any enhanced performance.

2.2 Use of Screenings to Produce HMA Mixtures Historically, many agencies have specified coarse-graded Superpave mixtures because it was thought that coarse-graded mixtures were less susceptible to rutting. This has led to a large amount of screenings that are not being utilized. In 2002, Cooley et al. (2) presented research on the use of screenings to produce HMA mixtures. The main objective of this study was to determine if rut resistant HMA mixtures could be attained with the aggregate portion of the mixture consisting solely of manufactured aggregate screenings (man-sand). Secondary objectives were to determine what effect both a modified asphalt binder and a fiber additive might have on rutting.

Two as-produced fine aggregates were used: a granite and a limestone. Table 2.1 shows the gradation for these aggregates. The limestone aggregate met current AASHTO gradation specifications for 4.75 mm NMAS mixtures. Two asphalt grades were used: PG 64-22 and PG 76-22. Mixtures were designed at three different Va (4, 5, and 6%) using 100 gyrations. There were eight mixture combinations of aggregate type, binder grade and fiber additive. The APA was used as a performance test to evaluate rutting potential.

5

TABLE 2.1 Gradations and Properties of Screenings (2)

Analysis of variance was used to evaluate the main factors affecting optimum asphalt content, VMA, %Gmm @ Nini, and APA rut depths. The factors that significantly affected optimum asphalt content were aggregate type, the existence of fibers, and design Va. Only two factors significantly affected VMA: aggregate type and the presence of fibers. %Gmm @ Nini was affected by aggregate type and design Va. Several factors affected APA rut depths: aggregate type, design Va and binder grade. Also, significant two- and three-factor interactions that affected rut resistance were 1) aggregate type with design Va, 2) aggregate type and binder grade, 3) fiber addition and design Va, 4) design Va and binder grade, 5) aggregate type, addition of fiber and binder grade, and 6) aggregate type, design Va and binder grade. The following conclusions were obtained from this research:

• Mixes having screenings as the sole aggregate portion can be successfully designed in the laboratory for some screenings but may be difficult for others.

• Screenings type, the existence of cellulose fiber, and design Va significantly affected optimum binder content. Of these three factors, screenings type had the largest impact on optimum binder content, followed by the presence of cellulose fiber and design Va, respectively.

• Screenings type and the presence of cellulose fiber significantly affected voids in mineral aggregate. Screenings material had a larger impact.

• Screenings type and design Va significantly affected %Gmm @ Nini results. Again, the screenings material had the largest impact.

• Screenings type, design Va, and binder type significantly affected laboratory rut depths. Of these three, binder type had the largest impact, followed by screenings type and design Va, respectively. Mixes containing a PG 76-22 binder had significantly lower rut depths than mixes containing a PG 64-22. Mixes designed at 4% Va had significantly higher rut depths than mixes designed at 5% or 6% Va.

6

Based upon the conclusions of the study, the following recommendations were provided:

• Mixes utilizing a screenings stockpile as the sole aggregate portion and having a

gradation that meets the requirements for 4.75 mm Superpave mixes should be designed in accordance with the recommended Superpave mix design system.

• Mixes utilizing a screenings stockpile as the sole aggregate portion but with gradations not meeting the requirements for 4.75 mm Superpave mixes should be designed using the following criteria:

Property Criteria

Design Air Void Content, % 4–6 Effective Volume of Binder, % 12 min. Voids Filled with Asphalt, % 67–80

2.3 Low Volume Road Applications Since the development of the Superpave mix design system, most placed Superpave designed asphalt mixtures have been designed for high traffic volume applications. One proposed use of 4.75 mm NMAS mixtures is for light traffic applications. Mixtures with 4.75 mm NMAS will generally have a surface with minimal surface voids, which creates a smooth, impermeable surface texture. These properties would be ideal in subdivisions and bike trials where there is high pedestrian and low vehicle traffic. Although the definition of a low volume road may differ between agencies, it may generally be considered as one with less than 1 million design Equivalent Single Axle Loads (ESALs).

Several states have successfully used 4.75 mm NMAS-like mixes for years. Alabama, Maryland, and Georgia have used these mixtures for thin overlays and preventative maintenance with good results. However, Superpave designed mixtures are not commonly used in low traffic applications throughout the U.S. This may be partly because some county and city agencies believe that costs of using Superpave mixtures are prohibitive. Also, there is concern that Superpave designed mixtures will result in lower optimum asphalt contents that will lead to reduced durability, which is important for a long-lasting, low volume mixture resistant to fatigue and thermal cracking. Since requirements for low volume roads may be quite different than their high volume counterparts, a literature review on Superpave designed mixtures for low volume applications is provided.

To determine if Superpave could be successfully utilized for low traffic volume applications, a number of agencies have carried out research to compare traditional Marshall designed mixtures with Superpave design methods (3,4,5). The general concern was that a Superpave designed mixture would adversely affect mixture durability with lower optimum asphalt content. Although different approaches were used by different agencies, researchers tried to determine the design gyration level that would provide asphalt contents and volumetrics similar to Marshall designed mixtures that have a good performance history. Prowell et al. (3)

7

found that a number of gyrations at design compaction (Ndes) of 68 gyrations provided designed binder contents similar to a 50 blow Marshall with optimum binder content selected at 6% Va. Mogawer et al. (4) recommends an Ndes of 50 gyrations for a low volume road in New England. Habib et al. (5) suggested that Ndes values used in Superpave mix design are about 20% higher than the required values. Habib concludes that lowering Ndes would result in increased asphalt contents for Superpave mixtures. Prowell and Habib both found that VFA Superpave requirements for these types of mixtures may be too restrictive.

The Iowa Highway Research Board (6) conducted a study of eight projects paved in 1998 to evaluate the performance of Superpave designed asphalt mixtures for low volume roads. Of the eight mixtures, three were 19 mm NMAS, four were 12.5 mm NMAS, and one was a 9.5 mm NMAS. All mixtures used a performance graded 58-22 binder. The objective of this research was to evaluate what issues affect the use of Superpave designed mixtures on low volume roads. Issues evaluated included economics, resources, and constructability. The final review of this research found that after six years, all the pavements constructed for this research exhibited excellent cracking resistance, except one project that had reflective cracks that began to appear a few weeks after placement. However, the authors did not relate that cracking to the use of Superpave designed mixtures but attributed it to the expected reflective cracking of a thin overlay on top of a PCC pavement. Rutting on all involved projects was well within the range of acceptable values, under 0.1 inch. The researchers found it impossible to get an objective measure of project costs compared to paving with conventional mixtures. However, the engineers and contractors involved in the projects believed that costs involved with the projects did not significantly increase.

In a 2004 article published in Asphalt Magazine (7), three county engineers were interviewed about their experiences with Superpave designed mixtures for low volume county roads. The interviews were from Blue Earth County, Minnesota; Stearns County, Minnesota; and St. Louis County, Missouri. All three county engineers found that Superpave was effective for county roads. However, Stearns County found that costs for using Superpave designed mixture on low volume roads were prohibitive, but this county still planned to use it on arterials and higher traffic roads. 2.4 Leveling and Patching Two possible uses for 4.75 mm NMAS Superpave mixtures are a leveling course or patching mix. A leveling course is defined as (8) a course (asphalt aggregate mixture) of variable thickness used to eliminate irregularities in the contour of an existing surface prior to superimposed treatment or construction. A smaller aggregate size mixture is beneficial for leveling applications where very thin lifts are needed to correct surface defects (9).

Patches are needed to repair weak areas in pavements, pot holes, or utility cuts. Structural patches should be designed and constructed with full depth asphalt concrete to ensure strength equal to or exceeding that of the surrounding pavement structure. Generally, there are three types of asphalt patching mixtures used (9): (a) hot mixed, hot laid, (b) hot mixed, cold laid, or (c) cold mixed, cold laid. Dense graded aggregates are used primarily for hot mixed, hot laid patching

8

mixtures. Typical gradations of dense graded patching mixtures are presented in Table 2.2, which shows that the current AASHTO gradation limits for 4.75 mm NMAS Superpave mixtures would fall within the limits of gradation C. The majority of all patching mixtures have 9.5 mm or 12.5 mm NMAS gradations (9). However, some agencies do specify a 4.75 mm NMAS mixture for patching. Larger NMAS mixtures seem to be preferred because they provide better stability, especially in deeper patches. When shallow holes are filled, a smaller NMAS mixture is beneficial, especially when the mixture must be feathered at the edges.

TABLE 2.2 Typical Gradations of Dense-Graded Patching Mixtures (9) Sieve Size Percent Passing

A B C 19.0 mm 100 12.5 mm 90-100 100 9.5 mm 75-90 90-100 100 4.75 mm 47-68 60-80 80-100 2.36 mm 35-52 35-65 65-100 1.18 mm 24-40 - 40-80 0.600 mm 14-30 - 20-65 0.300 mm 9-20 6-25 7-40 0.075 mm 2-9 2-10 2-10

2.5 Thin Overlays and Surface Mixtures It may be ideal to use 4.75 mm NMAS mixtures for thin overlays and surface mixtures. Hansen (10) stated that HMA overlays are probably the most versatile pavement prevention techniques available. They can improve structural capacity, improve ride, enhance skid resistance, reduce noise and improve drainage. However, thin overlays should only be placed on structurally sound pavements that exhibit surface distresses such as low-severity transverse and longitudinal cracking. According to NCHRP Report 531 (11), lift thickness should be at least three to four times the NMAS. For a thin overlay (less than one inch), a 4.75 mm NMAS asphalt mixture would meet a lift thickness to NMAS ratio of three to four.

The main function of a thin overlay of HMA may not be to provide strength to the pavement structure, but to protect a deteriorating pavement. If a thin overlay of 4.75 mm NMAS dense graded HMA is used as a surface mixture, it may provide a smooth, durable, water-tight surface. However, one possible concern of applying this type of mixture as surface mix is producing low surface texture. A low macro texture might lead to poor skid resistance, especially when a surface is wet.

9

2.6 NCAT Survey As part of the initial portion of this study, a survey of the current usages and possible future applications of this type of mix was sent to all US state highway agencies. Twenty-one states responded to the survey, as shown in Figure 2.1. Eleven of the 21 respondents did not utilize a 4.75 mm NMAS mix designation or similar mix type. Table 2.3 summarizes responses from agencies that have such a mix type.

FIGURE 2.1 Map of Respondents to NCAT Survey

10

TABLE 2.3 Summary of Responses for States Having a 4.75 mm-Like Mix

State Mix Design Method

Thickness or Spread Rate

In-place Density Requirement

Production is Expected To: Primary Uses

Arizona Arizona Method 50 lb/sy none decrease Surface mix

Delaware Superpave Varies none increase Leveling course

Georgia Superpave 85 lb/sy none N/A Leveling course

Illinois Superpave 3/4" 94% increase Leveling course

Missouri Marshall 1"–1.75" no remain steady Surface, leveling, overlay

North Carolina N/A 1" 85 or 90% remain steady N/A South Carolina Marshall 125 lb/sy none remain steady Surface mix

South Dakota Marshall 150 ton/mi none remain steady Leveling mix

Tennessee Marshall 35 lb/sy none remain steady Leveling mix

Washington N/A N/A N/A N/A N/A

West Virginia Marshall 70 lb/sy 92% increase Surface mix

Generally, three types of aggregates are used in these 4.75 mm mixtures: 1) rock or chip (0 to 30%), 2) screenings (0-50%); and 3) natural sand (0-30%). The most common grade of asphalt use was a performance grade 64-22. Hydrated lime mixed at 1% is commonly used as an additive; also mentioned was cement and liquid anti-strip additives. A large range of spread rates were reported; the average was 80 lb/sy, with the range being 35–125 lb/sy. Superpave and Marshall mix design methods are both used to design 4.75 mm NMAS asphalt mixtures; for Superpave mixtures an Ndes of 50 gyrations was typical. For the states that use Marshall designed mixtures, only Missouri disclosed the compaction effort used for their design (35 blows). Most states did not have current in-place density requirements for these mixtures; however three states do have in-place density requirements.

These mixture types were commonly used for leveling or scratch course, surface mixtures for low volume roads, and thin overlay for pavement maintenance. Appearance and performance were better than competing products and lower initial cost were cited as the most common advantages of this mixture type. Other listed advantages included the mixture’s ability to be placed in lifts less than one inch, relieve abundance of quarry fines, help retard reflective cracking, and reduce noise.

Generally, the disadvantages mentioned were that this type of mixture does not provide enough strength to the pavement structure and can be susceptible to rutting. Most states believed the production quantity would remain steady or increase over the next two years. The average production rate is about 420,000 tons per year. Individual responses for production rate are given in Table 2.4.

11

TABLE 2.4 Approximate Production of 4.75 mm NMAS Mixtures

The final question asked what aspects of this type of mixture should be further developed. States’ responses are given below:

• Florida: Leveling, thin overlays (maintenance/local agency) • New Jersey: Plan to use as leveling on a concrete pavement overlay on

an upcoming project. Right now we’re planning on using the 4.75 mm mix in AASHTO M323.

• Vermont: For low ESAL Superpave, it must be able to resist rutting and cold weather climate capabilities.

• Hawaii: Thin overlay for preventive maintenance • Nevada: Attempted to use a similar material in the past to fill

substantial cracking; after failed attempts and problems, use was discontinued.

• North Dakota: Bike trails • Washington: Thin wearing surfaces over structurally sound

pavement • Delaware: We are looking at the material for subdivision overlay

work. • Georgia: For low volume local roads, parking lots, etc. • Illinois: Explore ways to add macro texture to allow as a surface

course. • South Dakota: All types of roads (surface mix) • Missouri: Long-lasting surface mixtures for low volume

roadways • Iowa: Have an application as scratch course mix, but would not be specified for conventional HMA mixture (surface, intermediate, base)

Delaware <1000 tons Georgia: 320,000 tons for FY 2004 Illinois: Not yet adopted as common practice (N/A)

Tennessee: 225,000 tons West Virginia: 15,000 – 20,000 tons Arizona: 250,000 – 350,000 tons South Carolina: Low tonnage approximately 5% of total tonnage South Dakota: 75,000 tons Missouri: 1.7 million Surface level, and 750 thousand BP-2 North Carolina SF9.5A: 1,000,000 tons, S4.75A: 75,000 tons

12

An important finding from this survey was that 4.75 mm NMAS mixtures are being specified

and used as surface mixtures, leveling courses, and thin overlays. There appears to be benefits in using this type of mixture for these applications. Most states agreed that 4.75 mm NMAS mixes should be further developed to increase the mixture type’s overall structural capability and rutting resistance for use on low volume roads and in thin overlay applications.

13

CHAPTER 3 RESEARCH PLAN 3.0 RESEARCH PLAN In spring 2005, representatives from the eight participating states met at NCAT to determine the test plan for the 4.75 mm Superpave refinement pooled-fund study. Items discussed at this meeting included the following:

• Expected applications for 4.75 mm mixes in each state • Mix design criteria and concerns • Construction and performance concerns • Issues regarding specifics of performance testing (i.e., air void content for performance

testing, type of test used for durability testing, and load and tire pressure used for rut testing)

A comprehensive test plan was created from this meeting. The experimental test matrix is

shown in Table 3.1. This matrix shows that a 4.75 mm mix design was planned for all participating states using 50 gyrations and a design Va of 4%. Variations of those mix designs were planned by changing the design gyrations and the design Va. Additional variations were planned to evaluate changes in other mix factors such as dust content and binder grade. These are referred to as blend adjustment mixtures.

The first task was to obtain materials from each state. Participating states submitted a

proposed 4.75 mm blend representing the sources and general gradation from their state. Aggregate materials received from the participating states were tested to determine gradations and aggregate specific gravities. Alternate trial blends were then completed in addition to the blends submitted by the states. Thirteen aggregate blends from participating states were designed at 4% Va and 50 gyrations. Six of the 13 aggregate blends were also designed at 4% Va and 75 gyrations. An additional seven of the 13 aggregate blends were designed at 6% Va and 50 gyrations. Finally, three of the blends were designed at 6% Va and 75 gyrations. The 50 and 75 gyration compaction levels were selected because 4.75 mm mixes will likely be used for lower volume traffic applications (less than 3 million ESALs). Design Va of 4% and 6% were used to examine the concern of the mixes being over-asphalted due to high VMA values.

Also included in this study were four plant-produced baseline 4.75 mm mixtures from Mississippi, Maryland, Georgia, and Michigan that have known field performance and have been successfully used. These baseline mixtures served as benchmarks for comparing the results of the laboratory mix designs using the participating states’ materials.

The mix identification code used in this report to describe the mix designs is defined as

follows: The first two letters are used to define the state of origin (e.g., AL = Alabama). The first two numbers are the number of design gyrations, and the third number represents design Va (e.g., AL-50-6 = Alabama material designed at 50 gyrations and 6% Va). For blend adjustments, extra letters are given to describe the difference; for example, TNGM is used to denote Tennessee gravel mix, which is material from Tennessee but has a different source aggregate than the TN

14

mix design from Tennessee limestone. To describe blend adjustments—mixtures composed with the same material as in the original mix designs but prepared in different stockpile proportions—the letters “adj” have been attached (e.g., FL adj = Florida blend adjusted).

Table 3.1 Original Mix Test Matrix Ndes Gyrations 50 75 Air Voids 4.0 6.0 4.0 6.0 Aggregate Blend Florida FL-50-4 FL-75-4 Fl-75-6 Wisconsin WI-50-4 WI-50-6 Virginia VA-50-4 VA-75-4 Missouri MO-50-4 MO-50-6 Minnesota MN-50-4 MN-75-4 MN-75-6 Alabama AL-50-4 Al-50-6 Tennessee TN-50-4 TN-75-4 Connecticut CT-50-4 CT-50-6 New Hampshire NH-50-4 NH-75-4 NH-75-6 Blend Adjustment 1 WI adj-50-4 WI adj-50-6 Blend Adjustment 2 VA adj-50-4 VA adj-75-4 Blend Adjustment 3 FL adj-50-4 FL adj-75-6 Blend Adjustment 4 TNGM-50-4 TNGM-75-4 Baseline Mix 1 MS-50-4 Baseline Mix 2 GA-50-6 Baseline Mix 3 MD-75-3.5 Baseline Mix 4 MI-60-4

3.1 Test Methods 3.1.1 Aggregate Tests An aggregate analysis for gradation and specific gravity was performed on each virgin aggregate material sent to NCAT for this study. Gradations were performed in accordance with AASHTO (T 27), Sieve Analysis of Fine and Coarse Aggregate, and AASHTO (T 11), Materials Finer Than 75μm (No.200) Sieve in Mineral Aggregate by Washing. Specific gravities were determined by AASHTO (T 84), Specific Gravity and Absorption of Fine Aggregate. For the final blended aggregate determined from the mix design, AASHTO (T 304) Uncompacted Void Content of Fine Aggregate and AASHTO (T 176) Plastic Fines in Graded Aggregates and Soils by use of the Sand Equivalent Test were performed.

15

3.1.2 Mix Designs The AASHTO standard practice (R 35-3), Superpave Volumetric Design for Hot Mix Asphalt (HMA), was followed during the mix design phase of the study. This standard practice was used to verify specifications for 4.75 mm NMAS in AASHTO (M 323-04), Standard Specifications for Superpave Volumetric Mix Design.

Three aggregate blend gradations were selected for each of the eight participating state’s aggregate stockpiles. One of the three blends used in the aggregate trials was the blend proportion submitted by each state for their materials. The current gradation specification for 4.75 mm mixes is shown in Table 3.2; these control points were used in the blending process. Control points for the 4.75 mm sieve (90–100% passing) were strictly observed in the blending process to maintain a true 4.75 mm NMAS mix. However, some blend gradations were allowed to go outside of the control points on the #16 (1.18 mm) and #200 (0.075 mm) sieve so the effect of these limits could be evaluated. Also, since most states provided only two to three aggregate stockpiles, it was not always possible to develop reasonable alternative blends by proportioning the stockpile percentages and meet the current gradation limits. Figure 3.1 shows all the gradations used in this study plotted on a 0.45 power chart.

FIGURE 3.1 Gradations for State Mixtures

TABLE 3.2 4.75 mm Superpave Control Points

Sieve Minimum Maximum 12.5 100

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

sieve size

% Passing

0.075mm 1.180mm 4.75mm 9.5mm

16

9.5 95 100 4.75 90 100 1.18 30 60 0.075 6 12

Once three aggregate blends were determined, initial asphalt content was estimated for

each blend. Two replicate samples were prepared for each blend and mixed and conditioned in accordance with AASHTO R 30. Specimens were compacted in a Superpave Gyratory Compactor (Pine Instruments model AFG1A), following procedures in AASHTO T 312. This Superpave Gyratory Compactor was calibrated to provide an external angle of 1.25̊ . The internal angle, measured with a Pine model AFLS1 Rapid Angle Measurement kit, was 1.215 degrees. The bulk specific gravity of each compacted sample (Gmb) was determined by AASHTO T 166. Using AASHTO T 209, the theoretical maximum specific gravity of the asphalt mixture (Gmm) was determined for two samples of each blend. VMA, Va, VFA, D:B ratio, and %Gmm @ Nini were calculated for each trial blend. The volumetric properties of each blend were considered when selecting one of the three blends for the final mix design. In general, mixtures with the lowest estimate optimum asphalt content at the design Va were selected, as long as VMA, VFA, and D:B ratios were reasonable.

From the trial blend series, one mixture was selected for each state, and a binder series was run for the selected blend. In this part of the mix design process, three pairs of specimens were prepared and mixed at differing asphalt contents. The three asphalt contents were at the estimated optimum, at estimated optimum minus 0.5%, and at estimated optimum plus 0.5%. The volumetric properties of the mixtures were determined as mentioned above for the trial blend series, and a better estimate of the optimum asphalt at the desired Va was determined.

Finally, a set of two specimens was prepared with the selected aggregate blend and mixed at optimum asphalt content to verify the mix design. If the asphalt mixture compacted to the design Va and the volumetric properties were reasonable, the mix design was accepted for the study and samples were then prepared for performance tests. The 29 laboratory prepared mix designs are described in detail in Volume II: Mix Designs. 3.1.3 Performance Tests For each mix design and for baseline mixtures, a suite of performance tests were conducted. The performance tests were selected for analysis of permanent deformation, cracking resistance, permeability, and moisture sensitivity. For very thin lift applications and light-traffic pavements with low speed limits, rutting may not be a major concern. However, tests for permanent deformation were included to evaluate the stability of these mixes for other applications. Testing was conducted to evaluate volumetric criteria (e.g., VMA and VFA). Permeability tests were conducted to help evaluate possible in-place density requirements in the field. Testing was also performed on all the mixtures to evaluate their susceptibility to moisture damage.

17

3.1.3.1 Permanent Deformation. Permanent deformation testing was completed using a Mixture Verification Tester (MVT), shown in Figure 3.2. The MVT is a compact version of the APA. MVT testing followed AASHTO TP63-03, Rutting Susceptibility of Asphalt Pavements Using the Asphalt Pavement Analyzer. All specimens were tested using 100 lb wheel load and 100 psi hose pressure. For this study all specimen tests were conducted at 64°C. MVT specimens were tested at the design Va, 4.0% or 6.0%. Unlike the APA, the MVT only has the capability of testing two Superpave gyratory specimens or one beam specimen. The benefits of the MVT are it is smaller and lighter than the APA, which makes it more convenient for QC/QA applications in smaller laboratories. The MVT was used in this project since the amounts of material limited. The number of specimens required to perform the test was reduced from six to two by using the MVT.

FIGURE 3.2 Mixture Verification Tester

Research comparing MVT rutting to APA rutting is scarce. However, some work by Moore (12) at NCAT developed a correlation between the APA and MVT. Asphalt mixtures from the NCAT test track were used to compare the two devices, and it was found that the MVT generally had rut depths greater than those generated by the APA. This relationship is shown by Figure 3.3.

Mold

Loaded Wheel Pressurized Hose

18

R2 = 0.6819

0

2

4

6

8

10

12

14

0 2 4 6 8 10 12 14

MVT Rut Depth, mm

AP

A R

ut D

epth

, mm

Possible Outlier

FIGURE 3.3 APA Rut Depths Versus MVT Rut Depths

3.1.3.2 Moisture Damage Potential. Although there are several tests for assessing moisture susceptibility of asphalt mixes, the most commonly used is AASHTO T 283, Resistance of Compacted Asphalt Mixtures to Moisture-Induced Damage. AASHTO T 283 has been shown to be reasonably reliable and is commonly specified by most DOTs. A minimum tensile strength ratio (TSR) of 0.7 or 0.8 is typically used as the criteria. For this study, moisture susceptibility testing was performed following AASHTO T-283. At the panel meeting to discuss the testing plan for this study, representatives from the participating states decided that a higher Va should be used for some performance tests. AASHTO T-283 states that specimens should be compacted to 7+/-1% Va. The panel decided that the in-place Va after construction for 4.75 mm mixes would likely be in the range of to 8% to 10%. For this reason, specimens molded for moisture susceptibility in this study were targeted at 9 +/- 0.5% Va. 3.1.3.3 Permeability. In dense-graded asphalt pavements it is important to minimize permeability. Asphalt pavements with high permeability are susceptible to moisture damage and rapid aging. The factors that affect permeability are gradation, NMAS, and relative density. In-place density after compaction may be the most important factor influencing permeability. Previous studies at NCAT have shown that the critical in-place Va for permeability increases with smaller NMAS. As NMAS decreases, the size of the voids decrease, and thus, the interconnectivity of air voids decrease. This relationship was shown by Mallik et al. (13). Figure 3.4 shows permeability decreasing with smaller NMAS.

19

FIGURE 3.4 Best Fit Curves for In-Place Air Voids Versus

Permeability for Different NMAS (13)

Shape of the gradation curve is also an important factor that affects permeability. In general, coarse-graded mixtures have higher permeability than similar fine-graded mixtures, probably due to greater interconnectivity of voids in coarse-graded mixtures. Fine-graded mixtures tend to have smaller voids that are not as interconnected compared to coarse-graded mixtures of the NMAS.

Permeability testing for this research was accomplished using a falling head test (ASTM PS 121). This provisional standard is no longer used by ASTM; however it is similar to Florida Method (FM5-565). The target Va for the permeability test specimens was 9 +/- 0.5% for the same reason mentioned above. The specimens were compacted in a Pine Superpave Gyratory Compactor to a height of 55 mm and then saw-cut in half to obtain two samples about one-inch thick.

3.1.3.4 Cracking Resistance. There are several tests for fatigue cracking and thermal cracking. Fracture energy (FE) is one parameter that can be evaluated by indirect tensile (IDT) strength testing. Kim et al. (14) suggests that FE, which is the sum of strain energy and damage energy, may be a good indicator for the resistance of asphalt concrete to fatigue cracking. This claim is based on the observation that resistance of asphalt concrete to fatigue may be quantified by considering both resistance to deformation and resistance to damage. FE is obtained by integrating the area under the tensile stress-strain curve up to the point of fracture, shown in Figure 3.5. According to Birgisson et al. (15), fracture in a specimen is detected by monitoring the deformation differential and marking the location at which the deformation differential starts to deviate from a smooth curve; this is illustrated in Figure 3.6.

Kim et al. (14) compared several engineering IDT parameters measured on cores to observed fatigue performance data from Westrack. These parameters included 1) creep

20

compliance at 200 sec, 2) n-value, 3) IDT strength, 4) horizontal center strain at peak stress, and 5) FE. Of these five parameters, FE had the best correlation with the percentage of fatigue cracking. This relationship is seen in Figure 3.7. Kim suggests that based on this research, FE at 20oC is an excellent indicator of resistance of the mixture to fatigue cracking based on IDT testing of Westrack cores. Also, he proposed IDT testing at 20oC as a simple performance test for fatigue cracking.

FIGURE 3.5 Area Under Stress-Strain Curve at Point of Fracture (14)

21

FIGURE 3.6 Determination of Point of Fracture (15)

FIGURE 3.7 Relationship Between Field Fatigue Performance and Fracture Energy (14)

At this time, a method for determining FE has not been standardized. However, many

aspects of the test are found in AASHTO T 322-03, Strength of Hot Mix Asphalt (HMA) Using Indirect Tensile Test Device. FE testing was performed on an Instron Indirect Tension Tester (Figure 3.8) at 20oC, with a ram displacement rate of 50 mm per minute. Samples were molded

22

in a Superpave Gyratory Compactor (diameter = 150 mm) and then saw-cut on both sides to a height of 38 mm to 50 mm. Horizontal and vertical linear variable differential transducers (LVDTs) were mounted on both sides of the sample using a gauge length of 38.1 mm. Load was applied to the specimens until a peak load was reached and then began to decrease. A data acquisition system recorded load and LVDT data every 0.01 seconds. These data were then used to generate stress-strain curves. Procedures discussed in Fracture Energy from Indirect Tension Testing (14) were used in the calculation of FE density. FE density was computed as the area under the stress-strain curve to the point of fracture, illustrated in Figure 3.5. The point of fracture was determined by plotting the difference between the vertical and horizontal LVDTs on each side of the specimen. The point when the first side reached a maximum on this plot was taken as the time of fracture. This technique is illustrated in Figure 3.6. The equations and procedure for calculating FE are presented in Volume II.

Two sets of four samples were prepared for each mixture. Four samples remained unaged and four samples were long-term aged in a force-draft oven for six days at 85oC. . Both sets were tested for FE and then compared by calculating a ratio of unaged FE density to FE density after long-term aging.

FIGURE 3.8 Instron Indirect Tension Tester

IDT Specimen

LVDTs

Loading Head

23

CHAPTER 4 RESULTS AND ANALYSIS 4.0 RESULTS AND ANALYSIS

Twenty-nine mix designs were performed with aggregates from nine participating states. Details of the mix design development are included in Volume II. Table 4.1 summarizes the volumetric and aggregate properties of these mix designs.

TABLE 4.1 Mix Design Volumetric Properties State (mix) Va

(designtarget)

Ndes %A.C. VMA VFA %Gmm @ Nini

D:B Ratio

film thickness (microns)

SE FAA

AL-50-4 4.0 50 7.4 18.5 78.4 89.0 1.8 6.1 67 46.3 AL-50-6 6.0 50 6.9 18.8 68.1 87.2 2.0 5.4 67 46.3 TN-50-4 4.0 50 7.3 16.9 76.8 87.8 2.0 6.3 69 44.8 TN-75-4 4.0 75 6.8 16.0 74.8 87.2 2.2 5.7 69 44.8 MO-50-4 4.0 50 6.9 18.2 78.2 88.8 1.7 5.9 74 49.0 MO-50-6 6.0 50 6.2 18.4 66.7 86.9 2.0 5.1 74 49.0 VA-50-4 4.0 50 8.8 16.8 75.8 89.0 1.7 6.3 76 45.0 VA-75-4 4.0 75 8.3 15.8 74.9 88.5 1.9 5.8 76 45.0 FL-50-4 4.0 50 11.8 24.2 82.8 88.9 0.8 11.8 88 44.1 FL-75-4 4.0 75 11.0 22.6 81.8 88.4 0.9 10.8 88 44.1 FL-75-6 6.0 75 10.1 22.5 73.7 86.4 1.0 9.6 88 44.1 CT-50-4 4.0 50 8.8 19.9 80.9 86.6 1.2 8.9 79 40.7 CT-50-6 6.0 50 7.2 19.0 68.5 85.1 1.4 7.1 79 40.7 MN-50-4 4.0 50 8.8 21.1 80.4 87.5 1.6 7.4 67 46.2 MN-75-4 4.0 75 8.3 20.1 79.8 86.9 1.7 6.9 67 46.2 MN-75-6 6.0 75 7.4 19.7 70.1 85.3 1.9 5.8 67 46.2 NH-50-4 4.0 50 9.7 23.8 83.6 89.8 0.7 12.8 85 51.0 NH-75-4 4.0 75 9.3 22.9 84.0 89.4 0.7 12.1 85 51.0 NH-75-6 6.0 75 8.6 23.1 75.0 87.4 0.8 10.9 85 51.0 WI-50-4 4.0 50 7.5 18.0 77.4 87.7 1.2 8.9 81 43.7 WI-50-6 6.0 50 6.7 17.8 66.9 86.7 1.4 7.7 81 43.7

TNGM-50-4 4.0 50 9.7 20.9 80.7 88.1 1.0 9.2 70 42.2 TNGM-75-4 4.0 75 9.3 17.5 76.5 87.5 1.3 8.6 70 42.2 VA adj-50-4 4.0 50 9.0 16.8 76.4 88.5 1.7 6.5 76 45.0 VA adj-75-4 4.0 75 8.7 16.5 75.6 88.0 1.7 6.1 76 45.0 FL adj-50-4 4.0 50 10.0 20.6 81.1 88.9 1.7 7.9 79 44.5 FL adj-75-6 6.0 75 9.1 20.6 71.0 86.7 1.9 6.4 79 44.5 WI adj-50-4 4.0 50 6.8 16.1 74.4 87.1 1.9 6.8 81 45.8 WI adj-50-6 6.0 50 6.3 16.5 64.4 85.3 2.1 6.3 81 45.8

Table 4.2 shows a description of materials used for each state and stockpile percentages for each blend. Table 4.3 provides gradations used for each mixture and AASHTO gradation limits. Figure 3.1 is the gradation plot for all 12 aggregate blends.

24

TABLE 4.2 Materials and Stockpile Percentages for Laboratory Mixtures Stockpile 1 Stockpile 2 State (mix) Name Type % Name Type %

AL M-10 Granite 75 89s Granite 10 TN #10 hard Limestone 63 Natural Nat. sand 20 MO MO14 Dolomite 65 MO15 Dolomite 20 VA #10 Granite 75 Sand Nat. sand 25 FL Screenings Limestone 92 Sand Nat. sand 8 CT Stone Sand Trap rock 80 Screenings Trap rock 20 MN Minntac Tailings 87 Minntac fine Tailings 13 NH WMS Trap rock 69 D-Dust Trap rock 16 WI Man-sand Limestone 65 Screen 1/4" Limestone 20

TNGM # 10 Gravel 57 Sand Nat. sand 19 FLadj Screenings Limestone 91 Sand Nat. sand 3 WI adj Man-sand Limestone 56 Screen 1/4" Limestone 44

Stockpile 3 Stockpile 4 State (mix) Name Type % Name Type %

AL Shorter sand Nat.sand 15 TN #10 soft Limestone 17 MO MO13 Dolomite 15 NH RAP --- 15 WI Natural Nat. sand 15

TNGM #10 soft Limestone 18 Agg lime Agg lime 6 FLadj Fine Bag-house 6

25

TABLE 4.3 Blend Gradation for Laboratory Mixtures

Percent Passing

State (mix) 9.5 mm 4.75 mm 2.36 mm 1.18 mm 0.6 mm 0.3 mm 0.15 mm 0.075 mm AL 100.0 92.4 76.6 56.1 40.7 27.0 17.0 11.1 TN 100.0 94.4 69.1 48.7 33.8 19.0 13.8 11.6 MO 99.8 90.2 72.8 54.2 42.5 30.2 17.4 10.6

VA (1) 100.0 98.0 77.7 56.2 37.9 23.2 14.9 10.1 FL 100.0 95.6 78.8 57.7 41.0 26.0 11.7 7.7 CT 99.9 99.4 66.9 39.4 26.9 19.6 13.2 7.9 MN 100.0 98.0 86.4 61.1 38.6 23.1 14.8 11.2 NH 99.7 94.6 71.4 48.3 33.0 21.0 11.2 6.0 WI 100.0 90.8 63.1 41.5 26.7 14.9 9.4 7.1

TNGM 100.0 95.9 67.4 46.2 32.1 16.7 10.4 8.2 FL adj 100.0 95.6 79.1 58.1 41.9 29.0 17.1 13.4 WI adj 100.0 89.6 58.1 37.3 24.7 16.7 12.3 9.5

Note (1) – VA adj mixture used same gradation, but different binder grade and content 4.1 Mix Design Results 4.1.1 Optimum Asphalt Content Optimum asphalt content for the mixtures prepared in this study were relatively high compared to traditional Superpave designed mixtures. The average asphalt content for all 29 mixtures was 8.4% and the average effective asphalt content (Pbe) was 6.6%. The average asphalt absorption was 1.8%. FL-50-4 had the highest asphalt content and Pbe at 11.8% and 9.8%, respectively. MO-50-6 had the lowest optimum asphalt content at 6.2%. WI adj-50-6 had the lowest Pbe at 4.6%. The New Hampshire aggregate had the lowest asphalt absorption at 0.60%, whereas the Virginia aggregate had the highest amount of asphalt absorption at 3.6%.

26

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

AL 50-4

AL 50-6

TN 50-4

TN 75-4

MO

50-4

MO

50-6

VA 50-4

VA 75-4

FL 50-4

FL 75-4

FL 75-6

CT 50-4

CT 50-6

MN 50-4

MN 75-4

MN 75-6

NH 50-4

NH 75-4

NH 75-6

WI 50-4

WI 50-6

TNGM

50-4

TNGM

75-4

VA adj 50-4

VA adj 75-4

FL adj 50-4

FL adj 75-6

WI adj 50-4

WI adj 50-6

%AC

FIGURE 4.1 Optimum Asphalt Content

Figure 4.1 shows optimum asphalt content for each mix design. It can be seen that

increasing Ndes from 50 to 75 gyrations or increasing design Va from 4% to 6% will lower optimum asphalt content. The same trend for Pbe is illustrated in Figure 4.2. The statistical software package MINITAB was used to conduct an Analysis of Variance (ANOVA) to determine which design factors had a significant effect on Pbe. Three design factors were appropriate to use in this analysis: Ndes, design Va, and source material. Results of this analysis are shown in Table 4.4.

27

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

AL-50-4

AL- 50-6

TN-50-4

TN-75-4

MO

-50-4

MO

-50-6

VA-50-4

VA-75-4

FL-50-4

FL-75-4

FL-75-6

CT-50-4

CT-50-6

MN

-50-4

MN

-75-4

MN

-75-6

NH

-50-4

NH

-75-4

NH

-75-6

WI-50-4

WI-50-6

TNG

M-50-4

TNG

M-75-4

VA adj-50-4

VA adj-75-4

FL adj-50-4

FL adj-75-6

WI adj-50-4

WI adj-50-6

Effe

ctiv

e %

AC

FIGURE 4.2 Effective Asphalt Content

The ANOVA results for Pbe show that there is strong evidence to support the conclusion

that Ndes, design Va, and material source all influence asphalt content. Based on the F-statistics, material type has the largest influence on Pbe, followed by design Va.

TABLE 4.4 Analysis of Variance for Effective Asphalt Content

Source DF Seq SS Adj SS Adj MS F-stat P

Ndes 1 1.9892 0.8149 0.8149 29.93 0.000 Va(design) 1 2.9622 3.1157 3.1157 114.42 0.000 Material Source 15 48.3621 48.3621 3.2240 118.40 0.000 Error 14 0.3812 0.3812 0.0272 Total 31 53.6947

S = 0.165017 R-Sq = 99.29% R-Sq(adj) = 98.43%

To analyze the effect of design Va and Ndes, mix designs were separated into groups that had matching mix designs for each comparison. The comparisons were as follows:

• 50 gyrations (4% Va and 6% Va) • 4% Va (50 and 75 gyrations) • 75 gyrations (4% Va and 6% Va)

28

The mix design groupings are shown in Tables 4.5 through 4.7 and are used in comparison

evaluations in subsequent sections. Figures 4.3 through 4.5 show the mean asphalt contents for each grouping and the mean difference for each comparison. Figures 4.3 and 4.5 show that the difference in Pbe is 0.9% between 4% and 6% design Va for both compaction efforts. Figure 4.4 shows that the difference between mean asphalt content is 0.5% for 4% Va at 50 and 75 gyrations.

TABLE 4.5 Mix Design Comparisons for Ndes=50 (4%–6% Air Voids)

State IdAir voids (design) Ndesign %A.C. Eff AC% VMA VFA

% Gmm @ Nini DP SE FAA

film thickness (microns)

AL 4.0 50 7.4 6.30 18.5 78.4 89.0 1.8 67 46.3 6.1CT 4.0 50 8.8 6.80 19.9 80.9 86.6 1.2 79 40.7 8.9MO 4.0 50 6.9 6.10 18.2 78.2 88.8 1.7 74 49.0 5.9WI 4.0 50 7.5 6.00 18.0 77.4 87.7 1.2 81 43.7 8.9

WI2 4.0 50 6.8 5.1 16.1 74.4 87.1 1.9 81 45.8 6.8avg= 7.5 6.1 18.1 77.9 87.8 1.6 7.3

stdev = 0.8 0.6 1.4 2.3 1.0 0.3 1.5AL 6.0 50 6.9 5.60 18.8 68.1 87.2 2.0 67 46.3 5.4CT 6.0 50 7.2 5.50 19.0 68.5 85.1 1.4 79 40.7 7.1MO 6.0 50 6.2 5.30 18.4 66.7 86.9 2.0 74 49.0 5.1WI 6.0 50 6.7 5.20 17.8 66.9 86.7 1.4 81 43.7 7.7

WI2 6.0 50 6.3 4.6 16.5 64.4 85.3 2.1 81 45.8 6.3avg = 6.7 5.2 18.1 66.9 86.2 1.8 6.3

stdev = 0.4 0.4 1.0 1.6 1.0 0.3 1.1Diff = 0.8 0.8 0.0 10.9 1.6 -0.2 1.0

29

TABLE 4.6 Mix Design Comparisons for 4% Air Voids (50-75 Gyrations)

State IdAir voids (design) Ndesign %A.C. Eff AC% VMA VFA

% Gmm @ Nini DP SE FAA

film thickness (microns)

FL 4.0 50 11.8 9.70 24.2 82.8 88.9 0.8 88 44.1 11.8MN 4.0 50 8.8 7.20 21.1 80.4 87.5 1.6 67 46.2 7.4NH 4.0 50 9.7 9.10 23.8 83.6 89.8 0.7 85 51.0 12.8TN 4.0 50 7.3 5.80 16.9 76.8 87.8 2.0 69 44.8 6.3

TNGM 4.0 50 9.7 6.8 20.9 80.7 88.1 1.0 70 42.2 9.2VA 4.0 50 8.8 5.90 16.8 75.8 89.0 1.7 76 45.0 6.3VA2 4.0 50 9.0 6.0 16.8 76.4 88.5 1.7 76 45.0 6.5

avg = 9.3 7.2 20.1 79.5 88.5 1.4 8.6stdev = 1.4 1.6 3.3 3.2 0.8 0.5 2.7

FL 4.0 75 11.0 8.90 22.6 81.8 88.4 0.9 88 44.1 10.8MN 4.0 75 8.3 6.80 20.1 79.8 86.9 1.7 67 46.2 6.9NH 4.0 75 9.3 8.70 22.9 84.0 89.4 0.7 85 51.0 12.1TN 4.0 75 6.8 5.30 16.0 74.8 87.2 2.2 69 44.8 5.7

TNGM 4.0 75 9.3 6.4 17.5 76.5 87.5 1.3 70 42.2 8.6VA 4.0 75 8.3 5.40 15.8 74.9 88.5 1.9 76 45.0 5.8VA2 4.0 75 8.7 5.7 16.5 75.6 88.0 1.7 76 45.0 6.1

avg = 8.8 6.7 18.8 78.2 88.0 1.5 8.0stdev = 1.3 1.5 3.1 3.7 0.9 0.5 2.6Diff = 0.49 0.47 1.3 1.3 0.5 -0.1 0.6

TABLE 4.7 Mix Design Comparisons for Ndes=75 (4-6% Air Voids)

State IdAir voids (design) Ndesign %A.C. Eff AC% Binder VMA VFA

% Gmm @ Nini Dust ratio SE FAA

film thickness (microns)

FL 4.0 75 11.0 8.90 64-22 22.6 81.8 88.4 0.9 88 44.1 10.8MN 4.0 75 8.3 6.80 64-22 20.1 79.8 86.9 1.7 67 46.2 6.9NH 4.0 75 9.3 8.70 64-23 22.9 84.0 89.4 0.7 85 51.0 12.1

avg = 9.5 8.1 21.9 81.9 88.2 1.1 9.9stdev = 1.4 1.2 1.5 2.1 1.3 0.5 2.7

FL 6.0 75 10.1 8.00 64-22 22.5 73.7 86.4 1.0 88 44.1 9.6MN 6.0 75 7.4 5.80 64-22 19.7 70.1 85.3 1.9 67 46.2 5.8NH 6.0 75 8.6 7.90 64-24 23.1 75.0 87.4 0.8 85 51.0 10.9

avg = 8.7 7.2 21.8 72.9 86.4 1.2 8.8stdev = 1.4 1.2 1.8 2.5 1.1 0.6 2.7Diff = 0.8 0.9 0.1 8.9 1.9 -0.1 0.0 0.0 1.2

30

Air voids (design)

Mea

n of

Eff

AC%

6.04.0

6.0

5.0

4.0

5.2

6.1

Ndesign =50

FIGURE 4.3 Mean Effective Asphalt for 4% and 6% Air Voids (Ndes=50 )

Ndesign

Mea

n of

Eff

AC%

7550

8.0

7.0

6.0

5.0

4.0

3.0

6.7

7.2Design Air Voids = 4%

FIGURE 4.4 Mean Effective Asphalt Content for Ndes=50 and 75 (4% Air Voids)

31

Air voids (design)

Mea

n of

Eff

AC%

6.04.0

9.0

8.0

7.0

6.0

5.0

4.0

7.2

8.1

Ndesign =75

FIGURE 4.5 Mean Effective Asphalt Content for 4% and 6% Air Voids (Ndes=75 )

4.1.2 VMA The minimum VMA currently specified in AASHTO for 4.75 mm NMAS Superpave designed mixtures is 16.0%. Only one mixture (VA-75-4) barely failed to meet the current minimum VMA criterion. The average VMA was 19.3% for mix designs prepared for this research, and the maximum value was 24.2% (FL-50-4). Figure 4.6 shows all VMA values of the mix designs performed in this research.

32

14

16

18

20

22

24

26

AL 50-4

AL 50-6

TN 50-4

TN 75-4

MO

50-4

MO

50-6

VA 50-4

VA 75-4

FL 50-4

FL 75-4

FL 75-6

CT 50-4

CT 50-6

MN

50-4

MN

75-4

MN

75-6

NH

50-4

NH

75-4

NH

75-6

WI 50-4

WI 50-6

TNG

M 50-4

TNG

M 75-4

VA adj 50-4

VA adj 75-4

FL adj 50-4

FL adj 75-6

WI adj 50-4

WI adj 50-6

VMA

FIGURE 4.6 VMA Results for Each Mix Design

Figure 4.6 shows that the largest change in VMA occurs when the compaction level is

increased from 50 to 75 gyrations. This is expected since the aggregate will orient into tighter packing when the compaction energy is increased. As is well known, when designing asphalt mixtures, the addition of asphalt binder will decrease VMA until a minimum is reached, and any additional asphalt binder past this minimum will begin to push the aggregate structure open, increasing VMA. This effect explains why at a given Ndes, some mixtures slightly increase or decrease VMA as the optimum asphalt content decreases when the design Va changed from 4% to 6%.

To analyze the effect of Ndes, design Va, and material type on VMA, MINITAB was used to perform an analysis of variance. The results of this analysis are presented in Table 4.8. As with Pbe, the material type had the largest effect on VMA (F-stat = 44.4 and p-value = 0.000), and Ndes also had a significant effect (Fstat = 17.69 and p-value = 0.001) on VMA. Design Va, however, did not significantly influence VMA (F-stat = 0.05 and p-value =0.821).

33

TABLE 4.8 Analysis of Variance for VMA

Source DF Seq SS Adj SS Adj MS F P Source Material 12 170.023 171.551 14.296 44.44 0.000 Va (design) 1 0.430 0.017 0.017 0.05 0.821 Ndes 1 5.673 5.673 5.673 17.64 0.001 Error 14 4.503 4.503 0.322 Total 28 180.630

To illustrate the results of the analysis of variance, the comparison groupings in Tables

4.5 through 4.7 were used to show the differences in VMA due to changes in design Va and Ndes. Figure 4.7 shows that there is no difference in mean VMA for mixtures designed with 50 gyrations at 4% and 6% Va. For Ndes of 75, the mean difference in VMA is only 0.1% for 4% and 6% design Va, as shown in Figure 4.8. However, mixtures designed at 4% Va at 50 and 75 gyrations have a significant difference in mean VMA (1.3%), as illustrated in Figure 4.9.

Air voids (design)

Mea

n of

VM

A

6.04.0

20.0

19.0

18.0

17.0

16.0

15.0

14.0

18.118.1 Ndesign = 50

FIGURE 4.7 Mean VMA for 4% and 6% Air Voids (Ndes = 50)

34

Air voids (design)

Mea

n of

VM

A

6.04.0

22.0

21.0

20.0

19.0

18.0

17.0

16.0

21.821.9

Ndesign = 75

FIGURE 4.8 Mean VMA for 4% and 6% Air Voids (Ndes=75 )

Ndesign

Mea

n of

VM

A

7550

21

20.0

19.0

18.0

17.0

16.0

18.8

20.14 % Design Air Voids

FIGURE 4.9 Mean VMA for Ndes=50 and 75 at 4% Design Air Voids

4.1.3 VFA Three VFA ranges are currently specified in AASHTO M 323 for 4.75 mm NMAS mixtures, shown in Table 4.9. The average VFA for all mix designs in this study was 75.8. Only seven mix designs in this study met the strictest VFA criteria, which apply to mixes used on projects

35

with over 3 million ESALs. The maximum VFA observed was 84% for NH-75-4, and the minimum VFA was 64.4 for WI adj-50-6. Seventeen mix designs meet the VFA range for 0.3 to 3 million ESALs. Sixteen blends meet the VFA range for less than 0.3 million ESALs. Eight mixtures had VFA over 80%, and one was under 65%.

To analyze the effects of the design variables, an analysis of variance was performed using MINITAB. The results of this analysis are presented in Table 4.10. Design Va had the largest effect on VFA. Material source was also a significant factor. Ndes had the smallest influence on VFA (p-value = 0.118).

TABLE 4.9 AASHTO Specifications for 4.75 mm NMAS Superpave Mixtures

Design ESALs (Millions) Ndes

FAA Depth from Surface

SE VMA VFA Nini ≤ 100 mm ≥ 100 mm <0.3 50 - - 40 16.0 70-80% ≤91.5

0.3 to <3.0 75 40 40 40 16.0 65-78% ≤90.5 3.0 to<10 75 45 40 45 16.0 75-78% ≤89.0

Sieve size Min. Max. Va = 4.0% 12.5 mm 100 D:B Ratio: 0.9 to 2.0 9.5 mm 95 100 4.75 mm 90 100 1.18 mm 30 60 0.075 mm 6 12

TABLE 4.10 Analysis of Variance for VFA

Source DF Seq SS Adj SS Adj MS F P Source Material 12 282.136 246.366 20.530 18.45 0.000 Va(Design) 1 513.422 438.519 438.519 394.04 0.000 Ndes 1 3.083 3.083 3.083 2.77 0.118 Error 14 15.580 15.580 1.113 Total 28 814.221

S = 1.05493 R-Sq = 98.09% R-Sq(adj) = 96.17%

The comparison groups presented in Tables 4.5 through 4.7 illustrate the results of the

analysis of variance. Figure 4.10 shows the difference in VFA for mixtures with an Ndes= 50 at 4% and 6% design Va. Both groups had an average VMA of 18.1% for both design Va; the difference of 11.0% VFA is expected. Figure 4.11 shows a slight decrease in voids filled due to increasing Ndes from 50 to 75 gyrations at 4% Va. The mean difference in VMA for this comparison set was 1.3%. Figure 4.12 shows again the expected decrease in VFA by increasing design Va from 4% to 6% at 75 gyrations.

36

Air voids (design)

Mea

n of

VFA

6.04.0

80.0

70.0

60.0

50.0

40.0

66.9

77.9

Ndesign = 50

FIGURE 4.10 Mean VFA for 4% and 6% Air Voids (Ndes=50)

Ndesign

Mea

n of

VFA

7550

80

70

60

78.279.5

4% design airvoids

FIGURE 4.11 Mean VFA for Ndes= 50 and 75 (4% Air Voids)

37

Air voids (design)

Mea

n of

VFA

6.04.0

90.0

80.0

70.0

60.0

50.0

72.9

81.9

Ndesign = 75

FIGURE 4.12 Mean VFA for 4% and 6% Air Voids (Ndes=75)

4.1.4 Percent of Gmm at Ninitial Table 4.9 shows the current AASHTO requirements for relative density at Nini. For the two gyration levels evaluated (50 and 75), the corresponding Nini values are 6 and 7, respectively. Statistics for %Gmm @ Nini are provided in Table 4.11. All mixtures prepared for this research meet the specification limits for %Gmm @ Nini for the lowest two traffic levels. Two mixtures (NH-50-4 and NH-75-4) did not meet the most restrictive %Gmm @ Nini requirement of ≤89% for a design traffic level greater than 3 million ESALs.

TABLE 4.11 Descriptive Statistics for %Gmm @ Nini

Ndes N Mean Std Dev Minimum Median Maximum 50 18 87.7 1.3 85.1 87.8 89.8 75 11 87.4 1.1 85.3 87.4 89.4

The analysis of variance table, Table 4.12, shows that all three design factors had a significant effect on %Gmm @ Nini. The largest effect was due to changes in design Va. This is probably caused by a reduction in optimum asphalt content and the percent relative density required at Ndes when increasing design Va from 4% to 6%. Figures 4.13 through 4.15 show the differences in %Gmm @ Nini for the comparison groups. These comparisons show that increasing design Va had a substantial influence on %Gmm @ Nini. The average decrease in %Gmm @ Nini was 1.75% for both gyration levels when increasing design Va, whereas changing Ndes at 4% design Va was only 0.5%.

38

TABLE 4.12 Analysis of Variance for %Gmm @ Ninitial

Source DF Seq SS Adj SS Adj MS F P Ndes 1 0.5718 1.2686 1.2686 44.12 0.000 Va (design) 1 20.7127 12.9861 12.9861 451.66 0.000 Source Material 12 20.3516 20.3516 1.6960 58.99 0.000 Error 14 0.4025 0.4025 0.0288 Total 28 42.0386

S = 0.169563 R-Sq = 99.04% R-Sq(adj) = 98.08%

Air voids (design)

Mea

n of

% G

mm

@ N

ini

6.04.0

88.0

87.0

86.0

85.0

84.0

83.0

82.0

81.0

80.0

86.24

87.84

Ndesign =50

FIGURE 4.13 Mean %Gmm @ Nini for 4% and 6% Air Voids (Ndes=50)

39

Ndesign

Mea

n of

% G

mm

@ N

ini

7550

89.0

88.0

87.0

86.0

85.0

84.0

83.0

82.0

81.0

80.0

88.088.5

Design Air Voids = 4%

FIGURE 4.14 Mean %Gmm @ Nini for Ndes=50 and 75 (4% Air Voids)

Air voids (design)

Mea

n of

% G

mm

@ N

ini

6.04.0

89.0

88.0

87.0

86.0

85.0

84.0

83.0

82.0

81.0

80.0

86.4

88.2

Ndesign =75

FIGURE 4.15 Mean %Gmm @ Nini for 4% and 6% Air Voids (Ndes=75)

4.1.5 Dust-to-Binder Ratio and Film Thickness The D:B ratio range currently specified by AASHTO M 323 for 4.75 mm mixtures is 0.9 to 2.0. For the mix designs prepared in this study, the average was 1.5. The maximum was 2.2 for TN-75-4, and the minimum was 0.7 for NH-50-4 and NH-75-4. Two mixtures were above 2.0, and three were below 0.9. Since the D:B ratio is determined by dividing the percentage of dust by the

40

Pbe, it can be controlled by the asphalt and/or dust content of the mixture. Lowering asphalt content by increasing the design Va or Ndes will increase D:B ratio. From the analysis of variance table, Table 4.13, it is clear that changing design Va and gyration level have a significant influence of D:B ratio. Pbe is largely controlled by the gradation, and the percent of dust is a part of the gradation, so it was expected that the material source would have the largest influence on D:B ratio.

TABLE 4.13 Analysis of Variance for Dust-to-Binder Ratio

Source DF Seq SS Adj SS Adj MS F P Source Material 12 44.5793 44.7818 3.7318 137.05 0.000 Va (design) 1 4.6722 3.1157 3.1157 114.42 0.000 Ndes 1 0.8149 0.8149 0.8149 29.93 0.000 Error 14 0.3812 0.3812 0.0272 Total 28 50.4476

S = 0.165017 R-Sq = 99.24% R-Sq(adj) = 98.49%

Some researchers have proposed using film thickness (FT) requirements as an alternative

to specifying minimum and maximum values for VMA and VFA. FT was calculated for each mixture in this study. FT is simply the volume of effective asphalt divided by the estimated surface area of the aggregate. Surface area factors presented by Brown et al. (16) were used in this research for the calculation of FT. The average FT was 7.8 microns, the maximum was 12.8 for NH-50-4, and the minimum was 5.1 for MO-50-6. As with D:B ratio, all three experimental variables have an effect on FT, with material source having the largest influence. 4.1.6 Aggregate Properties (Gradation, SE, FAA) In the previous sections it was shown that mixture properties such as Pbe and VMA are primarily controlled by the source materials. Aggregate gradation is the most important factor in establishing the amount of voids created in the aggregate structure. Figure 4.16 shows that as VMA increases, the asphalt needed to fill voids increases. Since VMA and Pbe are both dependent on gradation, it was necessary to understand how gradation parameters influenced VMA of asphalt mixtures prepared for this study.

41

R2 = 0.9987

R2 = 0.9984

5.0

7.0

9.0

11.0

13.0

15.0

17.0

19.0

21.0

23.0

15.0 16.0 17.0 18.0 19.0 20.0 21.0 22.0 23.0 24.0 25.0

Voids in Mineral Aggregate

Volu

me

Effe

ctiv

e A

spha

lt

4% Design Air Voids

6% Design Air Voids

FIGURE 4.16 VMA versus Percent Volume Effective Asphalt

Fineness Modulus (FM) was calculated for each blend to examine the influence of gradation on VMA. The FM expresses how fine or coarse an aggregate blend is; the larger the FM, the coarser the gradation. To examine the effect of gradation on VMA, only the 13 mixtures designed at 50 gyrations and 4% Va were used to remove effects of compactive effort and different design Va.

Figure 4.17 shows two plots of FM versus VMA: one for mixtures with over 10% dust and one for mixtures with less than 10% dust. Since all the mixtures presented in this study were fine-graded, it was expected that coarser blends (i.e., higher FM) would have lower VMA, as seen in Figure 4.17 for both curves. Also, separating the mixtures into two groups (over and under 10% dust) showed that increasing dust content will lower VMA even for finer mixtures. Figure 4.18 illustrates that as the percent passing the 1.18 mm sieve increases, VMA increases. Again, the data are divided into two groups (over and under 10% dust), showing that VMA can be controlled with higher dust contents and/or adjusting the coarseness of the aggregate blend. However, using higher dust contents to control VMA can cause problems with other mix parameters such as higher D:B ratios and lower FT.

42

R2 = 0.3398

R2 = 0.4447

15.0

16.0

17.0

18.0

19.0

20.0

21.0

22.0

23.0

24.0

25.0

2.5 2.7 2.9 3.1 3.3 3.5 3.7

Fines Modulus

VMA

Under 10% Passing 0.075mm

Over 10% Passing 0.075 mm

FIGURE 4.17 Fineness Modulus Versus VMA for Over and Under 10% Passing the 0.075 mm Sieve

R2 = 0.2597

R2 = 0.5662

15.0

16.0

17.0

18.0

19.0

20.0

21.0

22.0

23.0

24.0

25.0

30 35 40 45 50 55 60 65

% Passing 1.18 mm Sieve

VMA

Under 10 % Passing 0.075 mm Sieve

Over 10% Passing 0.075mm Sieve

FIGURE 4.18 VMA Versus Percent Passing 1.18 mm Sieve for

Over and Under 10% Passing the 0.075 mm Sieve

43

The gradations for the mixtures are presented in Table 4.3 and plotted in Figure 3.1. Most mixes designed in this study have gradations considered to be fine-graded. The average percent passing the control sieves (4.75 mm, 1.18 mm, and 0.075 mm) was 94.9, 50.4, and 9.5, respectively. One mixture was below the 90% minimum percent passing the 4.75 mm sieve (WI-adj). This was the coarsest gradation, and it had one of the lowest VMAs in this research. One aggregate blend was over the 60% maximum percent passing the 1.18 mm sieve (MN). Even with a fairly high dust content of 11.2%, this blend had VMAs well above the 16% minimum. The blend adjustment from Florida (FL-adj) was the only aggregate blend outside the 6% to 12% range for percent passing the 0.075 mm sieve (P-075). In an attempt to reduce excessive VMA, 6% baghouse fines were added to the first Florida mix (FL) to create the FL-adj aggregate blend. By increasing the dust content to 13.4%, the VMA was reduced but was still relatively high at 20.6% for both FL-adj blends. This is probably due to the fine grading of the blend, (58.1% passing the 1.18 mm sieve and a FM of 2.792).

For mix designs below 0.3 million ESALs, there is currently no requirement for FAA. Between 0.3 to 3 million ESALs, the minimum FAA is 40. Mixes designed for greater than 3 million ESALs have a minimum FAA of 45 for mixtures used within 100 mm of the pavement surface, and minimum FAA of 40 for mixes used deeper than 100 mm from the pavement surface. The average FAA value was 45.2 for all the aggregate blends used as mix designs. The highest FAA was 51 for the New Hampshire blend, and the lowest was 40.7 for the Connecticut blend. Every blend met the 40 minimum FAA. Seven of the 13 blends met the 45 minimum FAA.

FAA did not significantly influence volumetric properties such as VMA, VFA, or Pbe. It has been thought that high FAA values probably increase VMA. For 4.75 mm NMAS mixtures, it seems that since 100% of the blend is fine aggregate, there would be a clear relationship between FAA and VMA, but this was not the case. Figure 4.19 shows no relationship between FAA and VMA. It also seems logical to assume that as FAA increases, the relative density at Nini would decrease because the more angular particles would create greater internal friction. However, the opposite trend was observed. Figure 4.20 shows that for the blends in this study, as FAA increased, so did the relative density at Nini. Since all the blends had FAA values above 40 and the average was 45.2, it is not possible to determine how blends with FAA below 40 would affect mixture properties and performance.

44

R2 = 0.0171

16.017.018.0

19.020.021.022.0

23.024.025.0

40 45 50 55

FAA

VMA

FIGURE 4.19 FAA Versus VMA for Ndes= 50 and Design Air Voids = 4%

R2 = 0.3889

86.0

86.5

87.0

87.5

88.0

88.5

89.0

89.5

90.0

40 42 44 46 48 50 52

FAA

Gm

m@

Nin

i

FIGURE 4.20 FAA Versus %Gmm @ Nini for Ndes=50 and Design Air Voids = 4%

For asphalt mixtures designed for over 3 million ESALs, the minimum SE value is 45;

for less than 3 million, the minimum is 40. All blends are well above these minimum values. The average was 76, the minimum was 67 for Alabama and Minnesota blends, and the maximum was 88 for the Florida blend. Since the amount of clay size particles relates to the amount of dust in the blend, SE is related to the amount of dust, D:B ratio, and FT. These relationships are shown in Table 4.14, where Pearson correlation coefficients and p-values are presented for each relationship. Since all the SE values for blends presented in this study are well above the minimum specified values, its effect on performance is not clear based on these results.

45

TABLE 4.14 Pearson Coefficients for Sand Equivalence

P-0.075 Dust-to-Binder Ratio Film Thickness

R -0.577 -0.570 0.679 p-value 0.039 0.042 0.011

4.2 Performance Tests 4.2.1 MVT Rut Depth The MVT was used to test permanent deformation on all 29 mixtures. The specimens used for this performance test were prepared at the design Va and compacted to Ndes. Since rutting on many mixtures was so severe, it was difficult to determine the effect of changes in air void, compaction level, and percent binder. All rut depths presented in this report were measured manually. Since the MVT device is programmed to shut off if the automatic rut depth measurements exceed 15 mm, many tests were automatically terminated before 8,000 cycles.

46

TABLE 4.15 Rut Depth and Mixture Properties for All Mix Designs

State (mix)

Va Ndes % Nat Sand

Pbe VMA VFA D:B SE FAA FT (microns)

Rut Depth (mm)

Cycles

AL 4.0 50 15 6.3 18.5 78.4 1.8 67 46.3 6.1 15.4 8000 AL 6.0 50 15 5.6 18.8 68.1 2.0 67 46.3 5.4 16.5 8000 TN 4.0 50 20 5.8 16.9 76.8 2.0 69 44.8 6.3 17.7 8000 TN 4.0 75 20 5.3 16.0 74.8 2.2 69 44.8 5.7 13.5 8000 MO 4.0 50 0 6.1 18.2 78.2 1.7 74 49.0 5.9 12.1 8000 MO 6.0 50 0 5.3 18.4 66.7 2.0 74 49.0 5.1 11.3 8000 VA 4.0 50 25 5.9 16.8 75.8 1.7 76 45.0 6.3 19.6 6228 VA 4.0 75 25 5.4 15.8 74.9 1.9 76 45.0 5.8 13.7 8000 FL 4.0 50 8 9.7 24.2 82.8 0.8 88 44.1 11.8 19.5 1205 FL 4.0 75 8 8.9 22.6 81.8 0.9 88 44.1 10.8 15.4 2047 FL 6.0 75 8 8.0 22.5 73.7 1.0 88 44.1 9.6 14.6 2425 CT 4.0 50 0 6.8 19.9 80.9 1.2 79 40.7 8.9 17.2 8000 CT 6.0 50 0 5.5 19.0 68.5 1.4 79 40.7 7.1 12.7 8000 MN 4.0 50 0 7.2 21.1 80.4 1.6 67 46.2 7.4 19.1 5724 MN 4.0 75 0 6.8 20.1 79.8 1.7 67 46.2 6.9 15.8 5256 MN 6.0 75 0 5.8 19.7 70.1 1.9 67 46.2 5.8 13.9 5074 NH 4.0 50 0 9.1 23.8 83.6 0.7 85 51.0 12.8 14.5 3595 NH 4.0 75 0 8.7 22.9 84.0 0.7 85 51.0 12.1 17.2 4220 NH 6.0 75 0 7.9 23.1 75.0 0.8 85 51.0 10.9 13.1 8000 WI 4.0 50 15 6.0 18.0 77.4 1.2 81 43.7 8.9 13.1 8000 WI 6.0 50 15 5.2 17.8 66.9 1.4 81 43.7 7.7 14.0 8000

TNGM 4.0 50 19 6.8 20.9 80.7 1.0 70 42.2 9.2 21.3 2795 TNGM 4.0 75 19 6.4 17.5 76.5 1.3 70 42.2 8.6 22.7 8000 VA-adj 4.0 50 0 6.0 16.8 76.4 1.7 76 45.0 6.5 9.8 8000 VA-adj 4.0 75 0 5.7 16.5 75.6 1.7 76 45.0 6.1 11.1 8000 FL-adj 4.0 50 3 7.9 20.6 81.1 1.7 79 44.5 7.9 14.3 8000 FL-adj 6.0 75 3 7.0 20.6 71.0 1.9 79 44.5 6.4 11.8 8000 WI-adj 4.0 50 0 5.1 16.1 74.4 1.9 81 45.8 6.8 5.3 8000 WI-adj 6.0 50 0 4.6 16.5 64.4 2.1 81 45.8 6.3 7.5 8000

Table 4.15 shows the rut depths for all 29 blends and the number of cycles the test performed before termination. The average rut depth was 13.3 mm for samples that completed 8,000 cycles. An interesting comparison is the average VMA for mixtures that completed 8,000 cycles to mixtures that did not complete 8000 cycles (see Figure 4.21). The average VMA for mixtures that completed 8,000 cycles was 18.2 mm and 21.5 mm for those mixtures terminated before 8,000. When VMA versus cycles to termination is plotted in Figure 4.22 for all mixtures, it is seen that for over 20% VMA, mixture rutting generally was so severe that the MVT device prematurely ended the test. There are some exceptions, one being VA-50-4, which had a relatively low VMA yet did not complete 8,000 cycles. This may be partly due to high percentage of natural sand (25%). The other exception is NH-75-6. This mixture had a high VMA (23.1%) yet completed 8,000 cycles and had a reasonable rut depth. This is probably explained by the mixture’s high FAA value of 51. NH-75-6 also had the lowest asphalt content for the three mixes prepared with the New Hampshire blend.

Based on the number of mixtures that did not complete a full 8,000 cycles on the MVT device, it is evident that limiting the VMA in 4.75 mm mixtures will be important in designing

47

rut-resistant mixtures. To analyze all the MVT data, including those mixtures that did not finish 8,000 cycles, rut depths were divided by the number of cycles completed for each mix to determine the total rutting rate in mm/cycle. When rutting rate is plotted against VMA, as shown in Figure 4.23, there are two separate trends for 4% and 6% Va. The 6% design air void line plots beneath the 4% air void line, and the lines diverge for higher VMA values. This indicates that even at higher VMA, the 6% air void mixtures were more rut resistant because of lower asphalt contents. This observation led to the consideration of evaluating the volume of effective asphalt (Vbe) as a parameter to control these mixtures. Vbe is simply VMA minus the Va and more directly quantifies the amount of binder needed for durability of a mix.

Figure 4.24 is a plot of the Vbe versus rutting rate, with the data sorted by the design Va. Figure 4.24 shows that the 6% and 4% air void curves are closer together than in Figure 4.23, which indicates that rutting for these laboratory mixtures is a function of the amount of asphalt, not just the total VMA.

Completed 8000 MVT Cycles

Mea

n of

VM

A

CompletedTerminated

23.0

22.0

21.0

20.0

19.0

18.0

17.0

16.0

15.0

18.2

21.5

Difference in mean VMA 3.3% ( p-value = 0.000 from 2 sample t test )

FIGURE 4.21 Mean Difference in VMA for Mixtures that Terminated Early and

Completed 8,000 Cycles on MVT

48

VMA

cycl

es

24.20

23.80

23.10

22.90

22.60

22.50

21.10

20.90

20.60

20.61

20.10

19.90

19.70

19.00

18.80

18.50

18.40

18.20

18.00

17.80

17.50

16.90

16.80

16.81

16.50

16.54

16.10

16.00

15.80

9000

8000

7000

6000

5000

4000

3000

2000

1000

0

Figure 4.22 VMA Versus Cycles to Termination

R2 = 0.6154

R2 = 0.35510.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

0.016

0.018

15.0 17.0 19.0 21.0 23.0 25.0

VMA

Rut

ting

Rat

e m

m/c

ycle

4% 6%

FIGURE 4.23 VMA Versus Rutting Rate by Design Air Voids

49

FIGURE 4.24 Volume of Effective Asphalt Versus Rutting Rate by Design Air Voids

When Vbe versus rut depth is plotted for all mixtures, Figure 4.25, the relationship is reasonable, with an R2 = 0.57. When the data is sorted in groups according to the amount of natural sand in each mixture, Figure 4.26, it is clear that as the percentage of natural sand increases, rutting rate also increases and the correlations improve. It appears that if Vbe is low, the effect of natural sand is minimized. However, if Vbe is over 13% to 14%, natural sand can be detrimental to rutting performance. The steep slope of the regression line for the over 15% sand mixtures seems to warrant limiting the amount of natural sand to less than 15% in mixtures designed for higher traffic volumes, where rutting resistance is important.

Figure 4.25 Vbe Versus Rutting Rate for All Mixtures Recall from Section 2.0 it was hypothesized that FAA may be an important indicator of a

mixture’s rutting resistance, since the majority of the aggregate in 4.75 mm mixtures passes the

50

4.75 mm sieve. Figure 4.27 shows Vbe versus rutting rate for mixtures with FAA over 45 and FAA under 45. For aggregate blends with FAA over 45, rutting rate increased with a linear relationship with increasing asphalt content. The curve is much steeper for aggregate blends with FAA less than 45. Figures 4.26 and 4.27 indicate that natural sand and FAA can influence a mixture’s rutting susceptibility, especially at asphalt contents over 14.0% by volume.

FIGURE 4.26 Vbe Versus Rutting Rate for all Mixtures, Sorted by Percent Natural Sand

R2 = 0.6705

R2 = 0.7378

0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

0.016

0.018

10.0 12.0 14.0 16.0 18.0 20.0 22.0

Volume of Effective Asphalt

Rut

ting

Rat

e m

m/c

ycle

<45 ≥45

FAA

FIGURE 4.27 Vbe Versus Rutting Rate for All Mixtures, Sorted by FAA

It has been shown that for the mixtures prepared in this study, asphalt content, percent

natural sand, and aggregate angularity all influence the rutting susceptibility of a 4.75 mm NMAS asphalt mixture. The question is, what is an acceptable amount of rutting for 4.75 mm

51

mixtures? Recall that Cooley et al. (1) recommended a maximum VMA of 18% on a limiting APA rut depth of 9.5 mm from NCHRP 9-17. Although the APA was not used in this research, a similar approach was used. Using the relationship shown in Figure 3.3, where MVT rut depths were correlated to APA rut depths, an equivalent MVT limiting rut depth was found to be 15.7 mm. However, the MVT tests were conducted with a hose pressure of 100 psi and wheel load of 100 lb. Since rut testing by Cooley (1) was conducted at 120 psi hose pressure and 120 lb wheel load, another problem exists in comparing the MVT data to the criteria. Using the correlation established by Prowell and Moore (12) at NCAT, shown in Figure 4.28, an equivalent critical rut depth for MVT was found to be 13.1 mm. This critical rut depth is easily converted to a critical rutting rate (13.1 mm = 0.00164 mm/cycle at 8000 cycles). Based on the regression shown in Figure 4.24 and the critical rutting rate of 0.00164 mm/cycle, the maximum Vbe should be 13.5%.

Based on 13.5% Vbe, the specified maximum VMA or VFA should be dependent on the design Va. For 4% design Va, the maximum VMA would be 17.5%, and a maximum VFA would be 77%. If a mixture were designed at 6% Va, the maximum VMA would be 19.5%, and the maximum VFA would be 69%.

y = 0.8972x - 0.8322R2 = 0.767

0

5

10

15

20

25

30

0 5 10 15 20 25 30

MVT Rut Depths at 120 lb / 120 psi, mm

MV

T R

ut D

epth

s at

100

lb /

100

psi,

mm

FIGURE 4.28 Relationship Between MVT Rut Depths at 120 lb, 120 psi to MVT Rut

Depths at 100 lb, 100 psi 4.2.2 Tensile Strength Ratio For all 29 mixtures designed in the laboratory, TSR was determined using AASHTO T-283. During a panel meeting of participating states, it was established that performance tests would be conducted at 9.0 percent% Va, since this is a likely in-place Va after compaction for a 4.75 mm NMAS mixture. Thus, all samples were compacted to 9±0.5% Va.

52

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

1.10

AL-5

0-4

AL-5

0-6

TN-5

0-4

TN-7

5-4

MO

-50-

4

MO

-50-

6

VA-5

0-4

VA-7

5-4

FL-5

0-4

FL-7

5-4

FL-7

5-6

CY-5

0-4

CT-5

0-6

MN-

50-4

MN-

75-4

MN-

75-6

NH-5

0-4

NH-7

5-4

NH-7

5-6

WI-5

0-4

WI-5

0-6

TNG

M-5

0-4

TNG

M-7

5-4

VAad

j-50-

4

VAad

j-75-

4

Flad

j-50-

4

Flad

j-75-

6

Wia

dj-5

0-4

Wia

dj-5

0-6

Tens

ile S

treng

th R

atio

FIGURE 4.29 Tensile Strength Ratios for 29 Mix Designs

Figure 4.29 shows a bar chart of TSR for all 29 mixtures. The average TSR was 0.65,

with a standard deviation of 0.19. The highest TSR was 0.99 for FL-adj-75-6, and the lowest was 0.23 for VA-50-6. Most agencies currently require a minimum TSR between 0.7–0.8. Only 12 of the 29 mixtures had TSR results greater than 0.7. However, it is important to note that the laboratory mixes did not contained any type of anti-stripping additive. Recall that the purpose of conducting the TSR testing in this study was to evaluate how changing the optimum asphalt content by adjusting Ndes or the design Va would affect the stripping potential.

It is also possible that low TSRs could have been caused by low vacuum pressures used to condition samples. It was noted during the saturation process of the conditioned samples that the vacuum pressure had to be reduced and saturation time generally had to be increased compared to other asphalt mixtures with larger NMAS. For 4.75 mm mixtures, the void spaces are smaller and not as interconnected and, therefore, it is possible that reducing the vacuum pressure and duration may have caused some damage to specimens by expanding void spaces and pushing apart aggregate. The low permeability results and the difficulty in obtaining saturation of specimens lends some evidence that 4.75 mm mixtures may be resistant to moisture intrusion, even at Va of 9.0%, and therefore, resistant to stripping.

The TSR results show that, in general, decreasing the asphalt content for most aggregate blends caused a slight increase in moisture damage susceptibility. However, several aggregate blends (TN, VA, NH, WI, and FL-adj) had an increase in TSR as asphalt content decreased. In Figure 4.30, dry and wet tensile strengths were plotted for each blend. It can be seen that for some source materials, dry strength increases with decreasing asphalt content. Other mixtures,

53

however, may show no difference in dry strength or even show a decrease in dry strength with lower asphalt contents. Asphalt-aggregate bonds are important to moisture susceptibility. This was not addressed in the experimental research plan for this study, so it is difficult to ascertain how the aggregate mineralogy affects the stripping potential of these mixtures.

0.0

50.0

100.0

150.0

200.0

250.0

AL-50-4

AL-50-6

TN-50-4

TN-75-4

MO

-50-4

MO

-50-6

VA-50-4

VA-75-4

FL-50-4

FL-75-4

FL-75-6

CY-50-4

CT-50-6

MN-50-4

MN-75-4

MN-75-6

NH-50-4

NH-75-4

NH-75-6

WI-50-4

WI-50-6

TNGM

-50-4

TNGM

-75-4

VAadj-50-4

VAadj-75-4

Fladj-50-4

Fladj-75-6

Wiadj-50-4

Wiadj-50-6

Tens

ile S

treng

th p

si

Dry Wet

FIGURE 4.30 Tensile Strengths for Conditioned and Unconditioned Samples

There is a weak relationship between VMA and TSR, as shown in Figure 4.31, and a

weak relationship between Vbe and TSR, as shown in Figure 4.32. Although these correlations are likely confounded by other variables, the general trend of increasing TSR with increasing VMA and effective asphalt content was expected.

Natural sand content is one factor that may affect TSR for these mixtures. Figure 4.33 shows a plot with only the mixtures designed at 50 gyrations and 4% Va to illustrate the influence natural sand has on sand asphalt mixtures. A plot with all the mix designs prepared for this research should show the same trend except that there would be more scatter around each point due to changes in asphalt content resulting from different gyration levels and target Va.

54

R2 = 0.2085

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

15.0 16.0 17.0 18.0 19.0 20.0 21.0 22.0 23.0 24.0 25.0

VMA

Ten

sile

Str

eng

th R

atio

FIGURE 4.31 VMA Versus Tensile Strength Ratio

R2 = 0.1722

0.00

0.20

0.40

0.60

0.80

1.00

1.20

10.0 12.0 14.0 16.0 18.0 20.0 22.0

Vbe

TSR

FIGURE 4.32 Effective Asphalt Content by Volume Versus Tensile Strength Ratio

55

R2 = 0.4517

0.000.100.200.300.400.500.600.700.800.901.00

0.00 5.00 10.00 15.00 20.00 25.00 30.00

Percent Natural Sand

Tens

ile S

tren

gth

Rat

io

FIGURE 4.33 Relationship with Percent Natural Sand in Blended Aggregate and TSR for

50 Gyration 4% Air Void Mix Designs

It was thought that FT may be a good indicator of TSR; however, for the blends in this study, the relationship was weak (R2=0.09). Dry strength seems to have a reasonable relationship with FT, as shown in Figure 4.34. No reasonable linear correlation or multiple linear regression models could be determined for wet tensile strength of the asphalt mixtures in this study.

R2 = 0.2378

0.00

50.00

100.00

150.00

200.00

250.00

4.0 6.0 8.0 10.0 12.0 14.0

Film Thickness

Dry

str

engt

h

FIGURE 4.34 Dry Strength Versus Film Thickness

56

4.2.3 Fracture Energy Density Ratio All 29 laboratory designed mixtures and three baseline mixtures were tested for FE density. For each mixture, two sets of specimens were prepared. The first set of specimens were tested with no aging. The second set was oven-aged at 85ºC for six days, then tested. A ratio of the aged FE density to the un-aged FE density was then calculated. A hypothesis of this study was that mixes with lower FE ratios would be more prone to aging and cracking over time. Although the aged FE values may not be below a threshold value where cracking will occur, a low ratio might identify a mixture that in certain field conditions could be more susceptible to cracking over time compared to mixtures with a higher ratio.

Table 4.16 shows FE ratios with aged and un-aged values for the 29 laboratory designed mixtures. The average FE ratio is 96.2%, with an average aged FE density of 5.399 kJ/m3 and an un-aged average of 5.574 kJ/m3. The high average for FE ratio indicates that small aggregate mixtures with high VMA and asphalt contents may be fairly resistant to cracking over time. FE ratios were plotted on Figure 4.35, sorted by state. In general, the FE ratio tends to decrease with decreasing asphalt contents that result from an increase in design Va and/or number of gyrations. For each source of materials, the 50 gyration and 4% air void mixture (50-4) had the highest asphalt content. However, several exceptions to decreasing ratio stand out (TN, MO, and VA), where the ratio increased with decreasing asphalt content. Figure 4.36 shows a weak relationship between Vbe and FE ratio, but there is a general trend of decreasing ratio with decreasing Vbe.

MINITAB was used to determine Pearson correlation coefficients between FE ratio and aggregate and mixture volumetric properties. The significant relationships are shown in Table 4.17. The strongest correlation with FE ratio was with D:B ratio. As can be seen in Figure 4.37, as D:B ratio increases, FE tends to decrease. Table 4.17 shows that there are also significant relationships between FE ratio and VMA, VFA, FT, and dust content. Figures 4.38, 4.39 and 4.40 show plots of FE ratio versus these properties. These relationships indicate that resistance to long-term cracking for 4.75 mm mixtures is affected to some degree by volume or mass proportions.

57

TABLE 4.16 Fracture Energy Results for Laboratory Mixtures

State (mix) P-075 Eff AC% VMA VFA D:B

Ratio FT

(microns) FE ratio

(%) FE unaged

(kJ/m3) FE aged (kJ/m3)

AL-50-4 11.1 6.30 18.5 78.4 1.8 6.1 102 3.57 3.65

AL 50-6 11.1 5.60 18.8 68.1 2.0 5.4 79 6.10 4.84

TN-50-4 11.6 5.80 16.9 76.8 2.0 6.3 60 3.70 2.20

TN-75-4 11.6 5.30 16.0 74.8 2.2 5.7 80 2.86 2.29

MO-50-4 10.6 6.10 18.2 78.2 1.7 5.9 59 5.84 3.45

MO-50-6 10.6 5.30 18.4 66.7 2.0 5.1 75 4.76 3.51

VA-50-4 10.1 5.90 16.8 75.8 1.7 6.3 68 4.54 3.07

VA-75-4 10.1 5.40 15.8 74.9 1.9 5.8 91 6.37 5.79

FL-50-4 7.7 9.70 24.2 82.8 0.8 11.8 127 4.50 5.72

FL-75-4 7.7 8.90 22.6 81.8 0.9 10.8 88 5.07 4.47

FL-75-6 7.7 8.00 22.5 73.7 1.0 9.6 94 5.67 5.35

CT-50-4 7.9 6.80 19.9 80.9 1.2 8.9 151 5.60 8.48

CT-50-6 7.9 5.50 19.0 68.5 1.4 7.1 104 6.90 7.15

MN-50-4 11.2 7.20 21.1 80.4 1.6 7.4 115 7.80 8.94

MN-75-4 11.2 6.80 20.1 79.8 1.7 6.9 110 7.38 8.08

MN-75-6 11.2 5.80 19.7 70.1 1.9 5.8 94 6.48 6.07

NH-50-4 6.0 9.10 23.8 83.6 0.7 12.8 137 5.45 7.45

NH-75-4 6.0 8.70 22.9 84.0 0.7 12.1 97 5.90 5.72

NH-75-6 6.0 7.90 23.1 75.0 0.8 10.9 106 7.06 7.48

WI-50-4 7.1 6.00 18.0 77.4 1.2 8.9 91 5.51 5.04

WI-50-6 7.1 5.20 17.8 66.9 1.4 7.7 85 6.05 5.17

TNGM-50-4 8.2 6.8 20.9 80.7 1.0 9.2 129 5.46 6.62

TNGM-75-4 8.2 6.4 17.5 76.5 1.3 8.6 97 5.06 4.89

VA adj-50-4 10.1 6.0 16.8 76.4 1.7 6.5 132 5.75 7.58

VA adj-75-4 10.1 5.7 16.5 75.6 1.7 6.1 93 5.69 5.32

FL adj-50-4 13.4 7.9 20.6 81.1 1.7 7.9 104 5.10 5.28

FL adj-75-6 13.4 7.0 20.6 71.0 1.9 6.4 60 5.89 3.55

WI adj-50-4 9.5 5.1 16.1 74.4 1.9 6.8 82 6.80 5.57

WI adj-50-6 9.5 4.6 16.5 64.4 2.1 6.3 81 4.79 3.86

Average 96.2 5.57 5.40

Std Dev 23.4 1.1 1.8

COV 24% 20% 33%

58

0

20

40

60

80

100

120

140

160

AL-50-4

AL 50-6

TN-50-4

TN-75-4

MO

-50-4

MO

-50-6

VA-50-4

VA-75-4

FL-50-4

FL-75-4

FL-75-6

CT-50-4

CT-50-6

MN

-50-4

MN

-75-4

MN

-75-6

NH

-50-4

NH

-75-4

NH

-75-6

WI-50-4

WI-50-6

TNG

M-50-4

TNG

M-75-4

VA adj-50-4

VA adj-75-4

FL adj-50-4

FL adj-75-6

WI adj-50-4

WI adj-50-6

Frac

ture

Ene

rgy

Rat

io

FIGURE 4.35 Fracture Energy Ratio for Laboratory Mixtures

y = 4.7625x + 26.252R2 = 0.2937

0.0

20.0

40.0

60.0

80.0

100.0

120.0

140.0

160.0

10.0 12.0 14.0 16.0 18.0 20.0 22.0

Volume of Effective Asphalt (%)

Frac

ture

Ene

rgy

Rat

io

FIGURE 4.36 Fracture Energy Ratio Versus Vbe

59

TABLE 4.17 Pearson Correlation Coefficients and p-Values for Linear Relationships with Fracture Energy Ratio

FT D:B VFA VMA P-075 R 0.532 -0.552 0.506 0.453 -0.418 p-value 0.003 0.002 0.005 0.013 0.024

y = -17.933x2 + 22.573x + 106.82R2 = 0.3233

0

20

40

60

80

100

120

140

160

0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3

Dust Proportion

Frac

ture

Ene

rgy

Rat

io

FIGURE 4.37 Fracture Energy Ratio Versus Dust-to-Binder Ratio

y = 4.1682x + 15.795R2 = 0.2056

0

20

40

60

80

100

120

140

160

15.0 17.0 19.0 21.0 23.0 25.0

VMA

Fra

ctu

re E

ner

gy

Rat

io

FIGURE 4.38 Fracture Energy Ratio Versus VMA

60

R2 = 0.3102

0

20

40

60

80

100

120

140

160

60.0 65.0 70.0 75.0 80.0 85.0 90.0

VFA

Frac

ture

Ene

rgy

Rat

io

FIGURE 4.39 Fracture Energy Ratio Versus VFA

R2 = 0.3061

0

20

40

60

80

100

120

140

160

4 6 8 10 12 14

Film Thickness (microns)

Frac

ture

Ene

rgy

Rat

io%

FIGURE 4.40 Fracture Energy Ratio Versus Film Thickness

A FE threshold was needed to discern critical values for volumetric properties such as

minimum VMA, VFA, or Vbe. Recall Figure 3.7 where Kim et al. (14) plotted FE density for specimens from Westrack. The regression from this plot indicates that fatigue cracking begins to occur at 3.0 kJ/m3. If this number is considered a threshold where no cracking is expected below the threshold value, then most of the mixtures presented in Table 4.16 should perform satisfactorily. However, this conclusion is not valid due to differences in field aging conditions at Westrack and the long-term oven aging used in the laboratory for this research.

61

Since an appropriate threshold value of this ratio is unclear, the FE ratios of the baseline mixtures presented in Table 4.18 were used as a benchmark to establish a reasonable limit for FE ratio. Due to a lack of material, FE density testing could not be performed for the baseline mixture from Mississippi. The mixtures from Georgia, Maryland, and Michigan are reported to have good in-service performance history. The mean FE ratio for the baseline mixtures is 76%, and the median is 80%. To serve as a benchmark for durability performance, the baseline median was chosen as a conservative estimate of a minimum value to compare with the laboratory prepared mixes. Figure 4.35 shows that only six mixtures were below the 80% FE ratio threshold.

The regression in Figure 4.35 shows that an 80% FE ratio corresponds to a minimum Vbe of 11.5, and Figure 4.37 shows that a minimum 80% FE ratio corresponds to a maximum D:B ratio of 2.0. Recommending only a minimum Vbe may not be sufficient with regard to assuring cracking resistance. It can be seen in Figures 4.37 and 4.40 that D:B ratio and FT have slightly stronger correlations with FE ratio compared to Vbe. Since FT and D:B ratio are both related to Vbe and dust content, it is clear that the ability to maintain cracking resistance for the 4.75 mm mixtures designed in this study is dependent on asphalt and dust contents. The currently specified maximum D:B ratio of 2.0 appears to be reasonable, based on Figure 4.37.

Table 4.18 Fracture Energy Density Data for Baseline Mixtures

State (mix) Va

(design) Ndes %A.C. Binder VMA VFA D:B

Ratio FT

(microns)

FE Ratio

%

Un-Aged FE

kJ/m3

Aged FE

kJ/m3 Mississippi 4.0 50 5.9 76-22 17.7 66.6 2.0 5.4 N/A N/A N/A Maryland 3.5 75 6.5 64-22 16.3 80.9 1.6 7.3 80 5.58 4.44 Georgia 6.0 50 6.0 64-22 16.7 76.4 1.5 6.7 81 4.89 3.95 Michigan 4.0 60 7.5 52-28 17.0 69.4 1.4 7.1 68 7.24 4.94 Mean = 76 5.90 4.44 Std Dev= 7.1 1.21 0.49 Median= 80 5.58 4.44 4.2.4 Permeability Laboratory permeability testing was performed on 27 of the 29 mix designs, and the results are shown in Table 4.19 and Figure 4.41. Mixtures TNGM-50-4 and VA-adj-50-4 were not tested for permeability due to insufficient material. Mixtures with permeability less than 125 cm/sec E-5 are generally considered impermeable, and Figure 4.40 shows that 21 out of the 27 mixtures are below this threshold. The maximum permeability was 211 E-5 cm/sec for WI-50-4, and the minimum was 8 E-5 cm/sec. It was thought that permeability would increase when asphalt content decreases; however, this was not the case.

It is clear that most of the mixtures prepared for this research are impermeable even at relatively high Va. It was mentioned in Section 2.3.3 that mixtures over 8.0% Va generally are permeable. The 4.75 mm NMAS mixtures were shown to be impermeable even at 9.0% Va because the Va are not as interconnected compared to larger NMAS asphalt mixtures.

TABLE 4.19 Permeability and Mixture Data for Laboratory Mixtures

62

State (mix) P-075 Pb VMA VFA D:B

Ratio SE FAA FT

(microns) k

(cm/s)E-5 AL-50-4 11.1 7.4 18.5 78.4 1.8 67 46.3 6.1 49 AL 50-6 11.1 6.9 18.8 68.1 2.0 67 46.3 5.4 50 TN-50-4 11.6 7.3 16.9 76.8 2.0 69 44.8 6.3 67 TN-75-4 11.6 6.8 16.0 74.8 2.2 69 44.8 5.7 75 MO-50-4 10.6 6.9 18.2 78.2 1.7 74 49.0 5.9 21 MO-50-6 10.6 6.2 18.4 66.7 2.0 74 49.0 5.1 38 VA-50-4 10.1 8.8 16.8 75.8 1.7 76 45.0 6.3 45 VA-75-4 10.1 8.3 15.8 74.9 1.9 76 45.0 5.8 30 FL-50-4 7.7 11.8 24.2 82.8 0.8 88 44.1 11.8 154 FL-75-4 7.7 11.0 22.6 81.8 0.9 88 44.1 10.8 126 FL-75-6 7.7 10.1 22.5 73.7 1.0 88 44.1 9.6 79 CT-50-4 7.9 8.8 19.9 80.9 1.2 79 40.7 8.9 91 CT-50-6 7.9 7.2 19.0 68.5 1.4 79 40.7 7.1 111 MN-50-4 11.2 8.8 21.1 80.4 1.6 67 46.2 7.4 8 MN-75-4 11.2 8.3 20.1 79.8 1.7 67 46.2 6.9 17 MN-75-6 11.2 7.4 19.7 70.1 1.9 67 46.2 5.8 8 NH-50-4 6.0 9.7 23.8 83.6 0.7 85 51.0 12.8 34 NH-75-4 6.0 9.3 22.9 84.0 0.7 85 51.0 12.1 52 NH-75-6 6.0 8.6 23.1 75.0 0.8 85 51.0 10.9 15 WI-50-4 7.1 7.5 18.0 77.4 1.2 81 43.7 8.9 211 WI-50-6 7.1 6.7 17.8 66.9 1.4 81 43.7 7.7 179 TNGM-75-4 8.2 9.3 17.5 76.5 1.3 70 42.2 8.6 162 VA adj-75-4 10.1 8.7 16.5 75.6 1.7 76 45.0 6.1 34 FL adj-50-4 13.4 10.0 20.6 81.1 1.7 79 44.5 7.9 177 FL adj-75-6 13.4 9.1 20.6 71.0 1.9 79 44.5 6.4 124 WI adj-50-4 9.5 6.8 16.1 74.4 1.9 81 45.8 6.8 31 WI adj-50-6 9.5 6.3 16.5 64.4 2.1 81 45.8 6.3 78 Average= 77 Std Dev= 59 COV= 77%

63

The results provided no clear relationships between mixture permeability and volumetric properties. One reason for this is that, according to the test procedure used (ASTM PS 121), a vacuum pressure of 525 mm of mercury (Hg) is to be applied to the specimen for five minutes to achieve saturation. However, due to low permeability of these mixtures, the specimens were saturated at a lower pressure (50–100 mm Hg) for 10 minutes until they were 85% to 95% saturated. This high level of saturation was used because it was observed that consistent results were only achieved at about 90% saturation. It is possible that the test specimens were damaged during the saturation process, which increased permeability due to expansion of internal voids.

FIGURE 4.41 Permeability for Laboratory Mixtures

One aggregate property found to influence mixture permeability is FAA. Figure 4.42 shows that permeability decreases with increasing FAA. Fine aggregates with high FAA can have flat and slivery particles, which in an uncompacted state like during the FAA test, results in high void contents. However, when the particles are compacted with a gyratory compactor, they become more horizontally oriented, which could block flow paths in the compacted specimens.

64

R2 = 0.36

0.0

50.0

100.0

150.0

200.0

250.0

40.0 41.0 42.0 43.0 44.0 45.0 46.0 47.0 48.0 49.0 50.0 51.0 52.0

FAA

Per

mea

bilit

y cm

/sec

E-5

FIGURE 4.42 FAA Versus Permeability

4.3 Baseline Mixtures Four plant-produced mixtures with good performance history were used as a baseline to compare the lab mixes to 4.75 mm NMAS mixtures being produced. Plant-produced mixtures from Mississippi, Maryland, Georgia, and Michigan were included as baseline mixtures. The mixture properties and averages are given in Table 4.18. The mixture from Georgia is not a 4.75 mm NMAS blend based on the percent passing the 4.75 mm sieve; however, it provides a good comparison to similar small aggregate size asphalt mixtures.

Generally, compared to the laboratory mixtures, baseline mixes are coarser graded, have lower optimum asphalt contents, lower VMAs, lower rut depths, higher TSR values, and lower average FE ratios. Figure 4.43 shows gradations for the baseline mixtures. Compared to the gradations of the lab mixes shown in Figure 3.2, the baseline mixtures are closer to the maximum density line. Even with lower percentages passing the 0.075 mm sieve, the baseline mixtures have lower VMAs due to coarser gradations.

The baseline mixture from Mississippi had the lowest MVT rut depth of all mixtures in the study. This was expected since this mix contained a polymer modified PG 76-22 binder. The average MVT rut depth for the baseline mixtures was 9.4 mm. This average is below the 13.1 mm rut depth assumed in this paper as a critical rut depth for 4.75 mm mixtures. The baseline mixture from Michigan had a 15.7 mm rut depth in the MVT, probably due to the use of a PG 58-22 binder in the mixture. Although this mix contained a softer asphalt grade, the MVT test was conducted at 64ºC, as were all mixtures in this study.

TABLE 4.20 Mixture Properties and Performance Data for Baseline Mixtures

65

State (mix) Va

(design) Va actual Ndes Passing 0.075

mm Passing 1.18

mm Passing 4.75

mm %Nat. sand

Mississippi 4.0 5.9 50 10.7 50.0 98.0 10.9 Maryland 3.5 3.1 75 8.1 42.8 95.6 15.0 Georgia 6.0 3.9 50 8.5 43.1 79.5 0.0 Michigan 4.0 5.2 60 7.1 54.6 92.5 0.0

Average= 4.4 4.5 58.8 8.6 47.6 91.4 6.5 Std Dev= 1.11 1.26 11.81 1.52 5.72 8.25 7.66

State (mix) %A.C. Pbe Binder VMA VFA %Gmm @ Nini D:B

Ratio Mississippi 5.9 5.3 76-22 17.7 66.6 86 2.0 Maryland 6.5 5.7 64-22 16.3 80.9 89.1 1.6 Georgia 6.0 5.5 64-22 16.7 76.4 90.2 1.5 Michigan 7.5 6.0 58-22 17 69.4 88.5 1.4

Average= 6.5 5.6 16.9 73.3 88.5 1.6 Std Dev= 0.73 0.30 0.59 6.52 1.78 0.26

State (mix) SE FAA FT

(microns) Rut Depth (mm) FE ratio % TSR k (cm/s)

E-5 Mississippi N/A N/A 5.4 3.8 N/A 0.85 48 Maryland 67 45.7 7.3 9.5 80 0.78 61 Georgia N/A N/A 6.7 8.6 81 0.92 107 Michigan 87 44.6 7.1 15.7 68 0.78 96

Average= 77.0 45.2 6.6 9.4 76.2 0.83 78 Std Dev= 14.14 0.78 0.85 4.90 7.08 0.07 28

State (mix) Dry TS Wet TS FE

Un-aged FE

Aged Mississippi 220.1 187.9 N/A N/A Maryland 164.4 129 5.582 4.442 Georgia 137.1 126.3 4.887 3.949 Michigan 209.4 164.1 7.242 4.937

Average= 182.8 151.8 5.904 4.443 Std Dev= 38.84 29.58 1.21 0.49

TSR for the baseline mixtures appear to be reasonable. The average was 0.83; however, if 0.80 is used as a minimum, which is common for many specifying agencies, the mixtures from Michigan and Maryland are slightly below this minimum. All baseline mixtures contained about 1.0 % hydrated lime, which may explain the higher TSR compared to the laboratory mixtures.

As with the laboratory designed mixtures, permeability was low even at high Va. The average permeability for baseline mixtures was 78 E-5cm/sec at 9.0% Va, which is practically the same as the average for the laboratory mixtures at 77 E-5 cm/sec at 9.0% Va.

66

4.75 mm Nominal Sieve Size

0.30

0.15

0.07

5

12.5

0

9.50

4.75

2.36

1.180.60

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

Sieve Size (mm)

Perc

ent P

assi

ng

Maryland

Mississippi

Georgia

Michigan

FIGURE 4.43 Gradations for Baseline Mixtures

One performance concern with these mixtures may be durability. FE ratios for baseline

mixtures are low compared to most of the laboratory mixtures. Additional aging caused by reheating the mixtures to make the specimens, lower FT, and lower Pbe probably contributed to the baseline mixtures’ lower FE ratios. Also, it is not clear if the softer binder used in the Michigan baseline mixture contributed to a lower FE ratio, which is noticeably lower at 68% compared to 80% and 81% for base line mixtures from Maryland and Georgia. 4.4 Review of AASHTO Specifications The AASHTO mix design criteria for 4.75 mm NMAS Superpave designed asphalt mixtures are presented in Table 4.21. The main objective of this research is to refine the current procedures and criteria for 4.75 mm mixtures, so a comparison to current AASHTO criteria is presented in this section.

67

TABLE 4.21 AASHTO Criteria For 4.75 mm NMAS Superpave Asphalt Mixtures.

Design ESALs (Millions) Ndes

Minimum FAA Depth from Surface Minimum

SE Min. VMA VFA

% Gmm @

Nini ≤ 100 mm ≥ 100 mm <0.3 50 - - 40 16.0 70-80% ≤91.5

0.3 to <3.0 75 40 40 40 16.0 65-78% ≤90.5 3.0 to<30 100 45 40 45 16.0 75-78% ≤89.0

Sieve size Min. Max. Va = 4.0% 12.5 mm 100 D:B Ratio: 0.9 to 2.0 9.5 mm 95 100 4.75 mm 90 100 1.18 mm 30 60

0.075 mm 6 12 4.4.1 AASHTO Gradation Limits Most of the laboratory prepared mixtures and baseline mixtures meet current gradation limits specified in AASHTO. Three blends, however, are outside current gradation limits. FL-adj had 13.4% passing the 0.075 mm sieve, which exceeded the maximum of 12%. This high dust content was intentionally used to lower the high VMA obtained in the FL blend. Six percent baghouse fines were added to the FL blend to lower VMA. For this mixture, adding the fines was beneficial. VMA lowered, TSR values increased, and D:B ratio increased to meet current specifications. This indicates that increasing the maximum limit on the 0.075 mm sieve may allow for 4.75 mm mix designs to have slightly higher dust contents as a way to control volumetric properties.

The MN blend was finer than the current limits specified for the 1.18 mm sieve. The maximum percent passing the 1.18 mm sieve is currently 60%; the MN blend had 61.1% passing. This gradation gave the lowest optimum asphalt content from the aggregate trial portion of the MN mix design. The final mixtures prepared with the MN aggregate blend had a high VMA (19.7 to 21.1), due largely to the fineness of the gradation.

The 1.18 mm sieve is used to divide a 4.75 mm NMAS mixture into two fractions, where the material above this sieve is the coarse fraction, and below the sieve is the fine fraction of the aggregate blend. Increasing the coarse fraction will make a fine-graded mixture move closer to the maximum density line. Figure 4.44 illustrates two ways that can be used to decrease Vbe. One way is to increase the dust content; the second way is to decrease the fine fraction of the gradation. It is recommended that the current gradation limits be adjusted to avoid gradations that may have excessive VMA. This can be done by limiting the amount of material passing the 1.18 mm to 55% and increasing the amount of material passing the 0.075 mm sieve to a maximum of 13.0%.

68

R2 = 0.47

R2 = 0.38

7.0

9.0

11.0

13.0

15.0

17.0

19.0

21.0

30.0 35.0 40.0 45.0 50.0 55.0 60.0 65.0

Percent Passing 1.180 mm Sieve

Vbe

Under 10 Percent

Over 10 Percent

Percent Pasaing 0.075mm Sieve

FIGURE 4.44 Vbe Versus Percent Passing 1.18 mm Sieve for

Over and Under 10% Passing the 0.075 mm Sieve 4.4.2 Sand Equivalent All aggregate blends in this study were well above the minimum specified limit for sand equivalence. The maximum SE result was 88 for the Florida blends; the minimum was 67 for the Minnesota and Alabama blends. The average for all the aggregate blends was 77. This study found no evidence to change the current AASHTO criteria for SE. 4.4.3 Dust-to-Binder Ratio As discussed in Section 4.1.5, five mix designs fell outside of the current specified range for D:B ratio. It was determined from the relationship shown in Figure 4.36 that the current specified maximum of 2.0 appears to be reasonable. However, the minimum D:B ratio may be slightly low. Figure 4.45 shows a plot of the average and median rutting rates for mixtures sorted by ranges of D:B ratio. It can be seen that higher D:B ratios tend to increase rutting resistance for these mixtures. In Section 4.2.1, a maximum allowable MVT rut depth was determined to be 13.1 mm, which is equivalent to a 0.0016 mm/cycle rutting rate at 8,000 cycles. The average D:B ratio for mixtures with a rutting rate of less than 0.0016 mm/cycle was 1.8 with only one mixture under 1.5 D:B ratio. Based on these data, it is recommended that the minimum D:B ratio be changed to 1.0, and for the ESAL range of over 3.0 million ESALs, a minimum of 1.5 is recommended.

69

FIGURE 4.45 Rutting Rate Versus Dust-to-Binder Ratio 4.4.4 Fine Aggregate Angularity Section 4.1.6 mentioned that there was no clear relationship between FAA and the volumetric properties of the mix designs prepared for this study. However, it was found that FAA did influence some of the results of the performance test. In Section 4.2.1 it was shown that an FAA over 45 reduced rutting at higher asphalt contents. Also, it was found in Section 4.2.4 that FAA over 45 may lower permeability. Based on these results, a FAA of over 45 may be appropriate for mixtures designed to higher ESAL ranges for both over and under a depth of 100 mm from the surface. 4.4.5 Percent of Gmm at Nini As mentioned in Section 4.1.4, only two mix designs failed to meet the most restrictive criteria for %Gmm @ Nini (≤89%). These mixtures also had relatively high rutting rates at 0.004 mm/cycle, which may indicate that they would be unstable when subjected to traffic. At this time there is no recommendation on modifying the current %Gmm @ Nini maximum. It was shown in Figures 4.13 and 4.15 that %Gmm @ Nini for 6% design air void mixtures averaged 1.7% lower than mixtures designed at 4% Va. However, if rutting rate is used as a measure of mixture stability, as shown in Figure 4.46, lowering the %Gmm @ Nini maximum for 6% design air void mixtures cannot be justified.

70

0.000

0.001

0.002

0.003

0.004

0.005

0.006

0.007

85.0 85.5 86.0 86.5 87.0 87.5 88.0

%Gmm@Nini

Rut

ting

Rat

e (m

m/c

ycle

)6.0 Percent Air Voids

FIGURE 4.46 %Gmm @ Nini Versus Rutting Rate for 6% Design Air Voids

4.4.6 Volumetric Requirements Currently, 4% Va is required by AASHTO M 323 for all NMAS mixtures. Results in Section 4.2 showed that 4.75 mm mixtures designed at 6% and 4% Va can have satisfactory performance test results. Relationships shown in Section 4.2.1 showed that mixtures designed at 6% Va have lower rutting than mixtures designed at 4% Va.

Most of the mixtures tested in this study, using aggregates from a wide variety of sources, had high VMA and, therefore, high asphalt contents. Rutting test results showed that rutting was more a function of Vbe than VMA. One way shown to reduce asphalt contents was to design these mixes at higher Va. This will allow mix designers to use existing aggregate materials that may yield blends with a high VMA, yet it will provide more reasonable and practical asphalt contents. For this reason, a range of design Va of 4% to 6% should be specified. Many of the mix designs in this study did not meet the current maximum VFA criteria of 78% for mix designs for over 0.3 million ESALs.

The three primary volumetric properties (Va, VMA, and VFA) are interrelated, and their criteria assure that mixtures have sufficient asphalt for durability but not too much asphalt that may lead to instability. The current criteria for 4.75 mm mixtures are to design the mixtures with 4% Va and a minimum VMA of 16%. The VFA criteria change depending on the traffic level. The current criteria can be restated as follows:

• Design Traffic < 0.3M ESALs: Vbe must be between 12.0% and 16.0%. • Design Traffic 0.3M to < 3.0M ESALs: Vbe must be between 12.0% and 14.1%. • Design Traffic 3.0M to < 30M ESALs: Vbe must be between 12.0% and 14.1%. Specifying a Vbe range is a more straightforward approach, since the limits to assure

durability and stability can be easily expressed with a single property. Based on Figure 4.35, a minimum Vbe of 11.5 was found to be appropriate based on the results of FE testing. Based on

71

Figure 4.27, a maximum Vbe of 13.5% is proposed for over 3.0 million design ESALs to limit the potential for rutting. 4.4.7 Summary of Recommendations for Revising 4.75 mm NMAS Mix Design Criteria Based on the analyses of the laboratory experiments, the following recommendations were proposed for revising the current AASHTO criteria for 4.75 mm mix designs:

• The target Va for selecting the design binder content should be changed to a 4.0% to 6.0% range. This will allow for a reduction in the design asphalt content for many 4.75 mm mixtures that have very high VMAs.

• VMA and VFA criteria should be replaced with minimum and maximum Vbe requirements. This is a more sensible approach when a range of design Va is used. For mixtures designed for projects less than 3 million design ESALs, a Vbe range of 12.0% to 15.0% is recommended. For mixtures designed for projects over 3 million ESALs, a minimum Vbe of 11.5% and a maximum Vbe of 13.5% are recommended. These limits were based on FE testing and MVT testing for the minimum and maximum Vbe, respectively.

• The maximum %Gmm @ Nini requirement appears appropriate for both 4% and 6% design Va. At this time, it is recommended that current Gmm @ Nini criteria be maintained.

• For aggregate blends designed for over 0.3 million ESALs, a FAA of 45 is recommended for improved rut resistance.

• For 4.75 mm NMAS asphalt mixtures designed for under 3.0 million ESALs, the minimum dust-to-binder ratio should be increased slightly, from 0.9 to 1.0. For mixtures designed for over 3.0 million ESALs, a minimum dust-to-binder ratio of 1.5 is recommended.

• The current maximum dust-to-binder ratio of 2.0 is appropriate based on the results of FE testing. It is recommended that the maximum dust-to-binder ratio of 2.0 be maintained.

• No evidence was found that suggested the current SE criteria should be adjusted. At this time, it is recommended that minimum SE criteria be maintained.

• The current gradation limits on the 1.18 mm and 0.075 mm sieve should be adjusted. Limits placed on percent passing the 1.18 sieve should be 30%–55%. Results of laboratory rutting tests showed that mixtures with gradations near the current control point of 60% passing the 1.18 mm sieve had severe rutting. Limits placed on P-075 should be 6.0% to 13.0%.

• It is recommended that 4.75 mm mixtures contain no more than 15% natural sand with an FAA under 45% to improve rutting resistance and moisture damage resistance, and to maintain low permeability.

From these recommendations, a summary of proposed mix design criteria is given in Table

4.22.

72

TABLE 4.22 Proposed Design Criteria for 4.75 mm NMAS Superpave Mixtures

Design ESAL Range (Millions) Ndes

Minimum FAA Minimum SE

Minimum Vbe

Maximum Vbe

%Gmm@Nini

D:B Ratio

<0.3 50 40 40 12.0 15.0 ≤91.5 1.0 to 2.0 0.3 to ≤ 3.0 75 45 40 12.0 15.0 ≤90.5 1.0 to 2.0 3.0 to ≤ 30 100 45 45 11.5 13.5 ≤89.0 1.5 to 2.0

Gradation Limits

Sieve Size Max. Min. Design Va Range = 4.0% to 6.0% 12.5 mm --- 100 9.5 mm 100 95 4.75 mm 100 90 1.18 mm 30 55

0.075 mm 13 6

73

CHAPTER 5 VALIDATION OF PROPOSED MIX CRITERIA BY PLANT PRODUCTION AND CONSTRUCTION 5.0 VALIDATION OF PROPOSED MIX CRITERIA BY PLANT PRODUCTION

AND CONSTRUCTION The proposal for this research stated that the 4.75 mm mix criteria would be validated by plant production and construction on a minimum of four projects. The field validation would examine the following issues:

• In-place densities after compaction • Appropriate spread rates and lift thicknesses • Workability of the mixture during construction • Variability in mixture volumetric and aggregate properties during production and

construction • Friction of in-place mixtures • Stability of the mixture during compaction • Permeability of in-place mixtures

The selection of the projects would take into consideration the four SHRP climate zones and

different levels of traffic. Each project was expected to have both a 4.75 mm mix and the normal mixture, so the evaluation of the 4.75 mm mix would be compared to a control mix. During the progress of the study it was difficult to select states specifically to fulfill the climate and traffic criteria. Pool-fund states were solicited to meet the minimum four-project criteria. The projects in this validation study were placed in Alabama, Missouri, Tennessee, and Minnesota. Only two of the four SHRP climate zones are represented: wet-freeze and wet-no freeze.

The production and placement of the 4.75 mm mixtures were generally independent maintenance projects, not tied to larger projects placing a control mixture. The agencies planned to compare the 4.75 mm mix field performance to other similar projects in the area, but similar projects were not specifically designated for comparison. The Minnesota project was also unique because it was a short research section on the MnRoad project, not a normal field construction project. The work plan for the validation spelled out specific tasks for the research team, as summarized below. (1) Assist the DOT with mix design and project specifications (2) During construction:

• Obtain plant-produced mix (both 4.75 mm mix and control mix) • Field lab test for Gmm and lab Gmb • Document production and construction • Test field compacted 4.75 mm pavement for permeability and friction • Obtain cores for density and lab testing

(3) Post-Construction Lab Testing: • Measure density of the cores

74

• Measure lab permeability of the cores • Extract and recover aggregate from plant-produced mix samples • Perform rut test by MVT or APA

In general, the field validation accomplished the tasks assigned. NCAT provided consultation

during the mix design process, but was not directly involved in laboratory mix designs for the mixtures placed. The NCAT laboratory prepared a number of preliminary mix designs for the Minnesota DOT, but the aggregates examined by NCAT were not used by the DOT. NCAT staff and the NCAT mobile laboratory were on site to collect sample and perform independent field density measurements. Permeability of the in-place 4.75 mm pavements was measured on the cores. Due to test equipment availability, complete surface friction measurements were obtained on only two projects. The post-construction testing in the NCAT laboratory included the test methods listed below:

• Lab test for Gmm – AASHTO T 209 • Lab test for Gmb – ASSHTO T166 • Lab procedure for binder content (by ignition) – AASHTO T308 • Lab test for gradation (washed) – AASHTO T 30 • Lab test for moisture susceptibility – AASHTO T 283 • Lab test for rutting (APA) – AASHTO TP 63 • Lab test for permeability – Florida Method FL 5-565 • Field test for friction using the Dynamic Friction Tester (DFT) – ASTM 1911 • Field test for friction using the Circular Texture Meter (CTM) – ASTM 2157

This section of the report includes a separate description of each of the four projects

(Sections 5.1 through 5.4) and a summary analysis (Section 5.5) comparing the field validation projects to the proposed mix design criteria. The analysis in this section uses the test results generated by NCAT to examine the 4.75 mm mix design, production, and placement. The comparison of the individual mix designs to the evolving mix design standards is intended to identify differences, not to imply a level of compliance or future performance. Predicted impact on short-term and long-term performance is solely based on commonly accepted HMA principles. 5.1 Alabama Field Validation Project 5.1.1 Project Description The ALDOT selected Wire Road just west of Auburn. The project number was STPNU-4423(200). The climate zone at this location is wet-no freeze. The traffic level for this section of Wire Road is 4700 AADT. Although there is insufficient information to estimate the design ESALs for this roadway, it is likely to be between 0.3 and 3.0 million ESALs. The plans called for placing the 4.75 mm HMA as a 0.75-inch surface lift. No conventional surface mix was placed as a control mix for performance comparison as part of this project.

75

The construction of the surface lift occurred from June 23 to June 26, 2006. The paving contractor was East Alabama Paving and the HMA production plant was located in Opelika. The mix was produced in a drum plant and paved with a conventional sequence of paving equipment. 5.1.2 Mix Design

The contractor prepared the mix design for the 4.75 mm HMA. The mix design was approved by the ALDOT on May 18, 2004. A copy of the mix design is included in the Appendix. Table 5.1 summarizes the approved mix design.

TABLE 5.1 Alabama Validation Project 4.75 mm Mix Design Summary Mix Type Proposed AASHTO Criteria Alabama 424 (surface mixture) Mix Size 4.75 mm NMAS 3/8-inch maximum aggregate size

(4.75 mm NMAS) Binder Type PG 67 -22 Binder Content 6.8 %, Pbe 6.53%,

Aggregate Blend

19% granite (#89 VMC Columbus, GA) 30% granite (M10 VMC Columbus, GA) 30% limestone (#8910 OCM Opelika) 20% man-sand (MM Pinkston Shorter) 1% baghouse fines

Target Gradation 30–55% passing 1.18 mm Sieve 6–13% passing 0.075 mm Sieve

47% passing 1.18 mm Sieve 6.0% passing 0.075 mm Sieve

Aggregate Properties FAA 45 (min) SE 40 (min) Nat.Sand 15(max) if FAA<45

FAA = 46 Not reported N/A

Air Voids 4.0–6.0% (Ndes=75 gyrations) 90.5 max (%Gmm @ Nini)

Va=3.3% at Ndes = 65 gyrations Nini = 89% of Gmm at 7 gyrations

Volumetric Properties Vbe 12.0 to 15.0 VMA 16.0 min (note 1) VFA 65-78 (note 1) D:B ratio 1.0-2.0

Vbe = 14.7 VMA = 18.0 VFA = 81.8 D:B ratio = 0.92

Moisture Susceptibility TSR = 0.85 with no anti-strip treatment Note 1 – current AASHTO criteria 5.1.3 Sampling and Testing Summary The 4.75 mm HMA was produced and placed over two days of paving. NCAT staff were on the project site to collect loose mix samples at the plant and locate and cut cores. Test equipment to measure initial in-place friction was not available during the placement of the 4.75 mm HMA. The loose production mix samples were transported back to the NCAT laboratory and compacted to measure field lab volumetric properties. Additional samples were taken back to the NCAT laboratory for extracted material proportions, moisture susceptibility, rutting, and permeability testing. Table 5.2 summarizes the production quality control test results performed by NCAT on the plant-produced mixture.

76

TABLE 5.2 NCAT Field Sampling and Testing for the Alabama Validation Project

Test [no. of samples / no. of replicates] Mix Design Target Production QC Mixture Va – Lab (%Gmm@Ndes) [4/3] 3.3% 2.2 – 3.4% Gmm [4/2] 2.467 2.444 – 2.482 Binder Content – by Ignition Method (Pb) [4/1] 6.8% 6.7 – 7.1%

Gradation – washed from ignition samples [4/1] 47% pass 1.18 6.0% pass 0.075

50.7 – 55.3 8.3 – 11.0

Vbe [4/3] VMA [4/3] VFA [4/3] D:B ratio

14.7 18.0 81.8 0.92

14.4 – 16.2 17.8-18.7 80.7-88.1

1.24 – 1.83 Moisture Susceptibility (TSR) [1/1] 0.85 0.80 Rut Testing – by MVT [1/2] 13.0 mm Lab Permeability from Field Cores (cm/sec) [3/1] 90 x 10-5 In-place Va – From Cores (note-1) [10/1] 11.7 avg, 9.5-13.2 Surface Friction – by DFT and CTM [0/0] Note-2 Note-1 Cores were taken at 200-ft intervals from Station 157+50 to 175+50. Note-2 DFT and CTM equipment were not available at the time of construction. 5.1.4 Analysis of Mix Design and Production Test Results

The binder content determined by the Ndes 65 gyration mix design procedure was 6.8%. The Pbe is 0.27% lower, indicating absorption by the aggregate. The range of binder contents of plant-produced mixture measured by the ignition oven was 6.7 to 7.1%. The consistency of the binder content is good and close to the mix design target.

The target gradation of the mix design was within the proposed control points for the 1.18 mm and 0.075 mm sieves. The gradation of the plant-produced mix was finer than the mix design target for both control sieves. The amount of aggregate passing the 1.18 mm sieve was 4% above the target. The amount of aggregate passing the 0.075 mm sieve was 3% above the target. The additional 3% dust content is very high. This production deviation from the mix design target is likely a factor in the observed low Va for lab compaction.

The mix design Va was 3.3%, which is below the recommended range of 4.0% to 6.0% for mix design. Lab voids from production mixture ranged from 1.9% to 3.7%, with an average of 2.8%. The mix design Va and, more importantly, the lab Va for production mix are lower than generally desired.

The proposed mix design criteria for Vbe is 12.0% –15.0% for mixes designed for 0.3 to 3.0 million ESALs. The mix design Vbe of 14.7% is within the proposed range. Several samples of the plant-produced mixes had Vbe results above the proposed mix design range of 15.0%. As noted in Phase I of this study, high Vbe can lead to an increase in rutting.

77

The VMA of the plant-produced mixture compacted in the lab to 65 gyrations was 17.8% to 18.7%. High VMA values coupled with low Va indicates that the mixture is supporting a high amount of asphalt binder. If the VMA does not collapse, the high binder content and low Va should slow down binder aging and, therefore, improve long-term durability.

The VFA criteria in the current AASHTO mix design specification for projects with between 0.3 and 3.0 million design ESALs is 65%–78%. The Alabama mix design exceeded the upper limit of the range. This also raises concern of the potential for rutting.

The 10 cores had an average measured in-place void content of 11.7%, with a range of 9.5% to 13.2%. These in-place air void results are well above the accepted norm of 8.0% for field compacted density for most types of dense-graded mixes, and even above the expected range for 4.75 mm NMAS mixtures of 8% to 10%. Permeability was measured in the laboratory on cores taken from the test section. Core permeability was 90 x 10-5 cm/sec. Even with the relatively high in-place Va, this mixture is impermeable. The measured lift thickness was 16.8 to 21.9 mm and is comparable to the intended thickness of 19 mm. Moisture sensitivity of the plant-produced mix had a TSR of 0.80 based on IDT tests of laboratory-prepared specimens at 9.0% Va. The average tensile strength was 114.5 for the unconditioned specimens and 91.9 psi for conditioned specimens. Although the field TSR was lower than the mix design TSR, it is probable that the lower computed TSR was impacted by the high Va of the specimens.

Rutting of the plant-produced 4.75 mm HMA with the Mix Verification Tester resulted in a rut depth of 13 mm at 8,000 cycles. This level of rutting is consistent with the 12 to 17 mm MVT rutting presented in Section 4.2.1 from the laboratory evaluation. 5.2 Missouri Field Validation Project 5.2.1 Project Description The Missouri DOT selected Dunklin County Route EE, near Kennett, from State Route 153 to State Route 25. The climate zone at this location is wet-freeze. The traffic level for this test section is 2,500 AADT and less than 5% trucks. The design traffic for this project was assumed to fall in the lowest traffic category: less than 0.3 million ESALs. The plans called for placing the 4.75 mm HMA as a 0.75-inch surface lift. No conventional surface mix was placed as a control for performance comparison as part of this project.

The construction of the surface lift occurred on August 16, 2007. The paving contractor was Apex Paving Company, and the HMA was produced at the Delta Asphalt Plant. The mix was produced in a drum plant and paved with a conventional sequence of paving equipment.

78

5.2.2 Mix Design

The contractor prepared the mix design, and the MoDOT approved the design on April 26, 2007. A copy of the mix design is included in the Appendix. Table 5.3 summarizes the approved mix design.

TABLE 5.3 Missouri Validation Project 4.75 mm Mix Design Summary Mix Type Proposed AASHTO Criteria Missouri BP-3 Plant Mix Bituminous Mix Size 4.75 mm NMAS 4.75 mm NMAS Binder Type PG 64-22 Binder Content 6.4%, Pbe=5.4%

Aggregate Blend 55% dolomite (LD Williamsville #1) 25% man-sand (MSGV BS&G Dexter) 20% nat-sand (NS1 BS&G Dexter, MO)

Target Gradation 30–55% passing 1.18 mm Sieve 6–13% passing 0.075 mm Sieve

48% passing 1.18 mm Sieve 7.6% passing 0.075 mm Sieve

Aggregate Properties FAA = 40 (min) SE = 40 (min) Nat.Sand=15(max) if FAA<45

FAA = 45 Not reported N/A

Air Voids 4.0–6.0% (Ndes=50 gyrations) 91.5 max (%Gmm @ Nini)

Va = 4.0% at Ndes = 50 gyrations Not reported

Volumetric Properties Vbe 12.0-15.0 VMA 16.0 min. (note 1) VFA 70-80 (note 1) D:B ratio 1.0-2.0

Vbe = 12.2 VMA = 16.3 VFA = 75.2 D:B ratio = 1.4

Moisture Susceptibility Not tested, generally not required for

mixtures on low volume roads Note 1 – current AASHTO criteria 5.2.3 Sampling and Testing Summary

The 4.75 mm HMA was produced during a single day of paving. NCAT staff was on the project site with a mobile laboratory to collect loose production mix samples at the plant, compact production mixture, measure field-lab volumetric properties, and locate and cut cores. The CTM was used to measure surface macrotexture, but the DFT equipment to measure initial in-place friction was not available during the placement of the 4.75 mm HMA. Samples were taken back to the NCAT laboratory for extracted material proportions, moisture susceptibility, rutting and permeability testing. Table 5.4 summarizes the production quality control test results performed by NCAT on the plant-produced mixture.

79

5.2.4 Analysis of Mix Design and Production Test Results

The design binder content for the 50-gyration mix design procedure was 6.4%. The Pbe is 1.0% lower, indicating a high amount of absorption by the dolomite aggregate. The range of binder contents of the plant-produced mixture measured by the ignition oven was 6.83% to 7.42%. The plant-produced binder contents were consistently about 0.5% above the mix design target.

The target gradation of the mix design was within proposed control points for the 1.18 mm and 0.075 mm sieves. The gradation of the plant-produced mix was finer than the mix design target for both control sieves. The NCAT-extracted gradations from production mix showed a wide production range, 49% to 58% passing the 1.18 mm sieve. The range of NCAT results for P-075 were very consistent (11.8–12.3), but was 4.5% above the mix design target. However, the dust contents of the plant mix were still within the proposed mix design gradation control point. The large difference between the mix design dust content and the results from production mix tests obtained by NCAT can be partly attributed to differences between dry gradations and washed gradations. NCAT used washed gradation analyses, whereas MoDOT uses dry gradations. Higher dust contents may also be due to breakdown of the gradation in the plant and breakdown in the ignition oven with the dolomite aggregate.

TABLE 5.4 NCAT Field Sampling and Testing for the Missouri Validation Project Test [no. of samples / no. of replicates] Mix Design Target Production QC Mixture Va – Lab (%Gmm@Ndes) [3/6] 4.0% 3.6 – 4.9% Gmm [3/2] 2.456 2.453 – 2.460 Binder Content – by Ignition Method (Pb) [3/1] 6.4% 6.8 – 7.4% Gradation – Washed from Ignition Samples [3/2] 48 pass 1.18

7.6 pass 0.075 49 - 58

11.8 – 12.3 Vbe VMA VFA D:B ratio

12.2 16.3 75.2 1.4

12.5 – 13.3 16.6 – 17.7 74.3 – 76.5 2.1 – 2.2

Moisture Susceptibility (TSR) [3/1] Not tested 0.66 – 0.74 Rut Testing – by APA [3/4,6 note-2] 6.7 mm Lab Permeability from Field Cores (cm/sec) [10/1] 40 x 10-5 In-place Va – from Cores [10/1] 10.1 avg, 9.2 – 11.9 Surface Friction – by DFT(note-1) and CTM [10/2] MPD 0.17- 0.22 mm Note-1 The DFT was not available for this project. Note-2 There were 4 replicates for sample 1 and 6 replicates for samples 2 and 3.

For the MoDOT 4.75 mm HMA, the lab voids from production mix ranged from 3.6% to 4.9%. This range is within a normal specification production tolerance of +/- 1.0%. Field notes indicate that the aggregate proportions were changed to increase the man-sand and decrease the natural sand after the contractor’s initial production mix QC sample measured 3.3% Va. Based on the range of lab Va on production mix measured by NCAT, the aggregate proportion change improved the aggregate structure of the mixture to maintain the target 4.0% voids.

80

The mix design Vbe of 12.2% and the Vbe results for the plant-produced mixture were

within the proposed specification range of 12.0%–15.0%. The mix was designed with a VMA of 16.3%, and the VMA from plant-produced mixture was 16.6% to 17.7%. The higher VMA for the production mix compared to the mix design was sufficient to accommodate the additional asphalt binder and mineral filler without reducing the Va. The VFA of the mix design and the plant-produced mix were also within the current mix design specification range of 65–78.

The ten cores had an average measured in-place void content of 10.1%, with a range of 9.2% to 11.9%. Permeability tests on the cores resulted in an average permeability of 40 x 10-5 cm/sec, which confirmed that the 4.75 mm layer was impermeable. The measured lift thickness was 19.4 to 24.5 mm and is at or above the intended thickness of 19 mm.

A moisture sensitivity test was not reported for the mix design. Moisture sensitivity of the plant-produced mix yielded an average TSR of 0.70 with a range of 0.66 to 0.74, based on IDT tests of laboratory-prepared specimens ranging from 6.5% to 7.6% Va. The TSR value was based on an average conditioned tensile strength of 102 psi and an average unconditioned tensile strength of 145 psi.

Rutting of the plant-produced mix compacted in the laboratory to 4.0% Va and tested with the APA resulted in an average rut depth of 6.7 mm at 8,000 cycles. Based on the correlation of APA and MVT results presented in Figure 3.3, this result would be approximately 11 mm in the MVT, which is good compared to the 12–17 mm range for MVT results from the laboratory evaluation.

The CTM was used to measure the macrotexture of the 4.75 mm HMA surface after compaction. Readings were taken at each of the core locations. The mean profile depth (MPD) measured from 0.17 to 0.22 mm. This texture measure is normal for a fine-graded dense HMA with a small NMAS. The DFT was not in service at the time of this construction. No DFT measurements were taken. 5.3 Tennessee Field Validation Project 5.3.1 Project Description

The Tennessee DOT selected State Route SR 25 in Robertson County. The project number was 74000-4200-404. The project began at log mile 4.00 and ended at log mile 6.95. The climate zone at this location is wet-no freeze. The traffic level for this section of SR 25 was 1620 AADT with 18% trucks. The design traffic for this project was assumed to fall in the second-lowest traffic category: 0.3 to 3.0 million ESALs. The posted speed limit was 55 mph. Predominant distress in the existing HMA pavement was transverse cracking at 10 to 40-ft spacing. The plans called for placing the 4.75 mm HMA as a 0.75-inch surface lift. Two 4.75 mm mixes were placed, a virgin mix was placed in the east-bound lanes, and a mix with 15% reclaimed asphalt pavement (RAP) was placed in the west-bound lanes. No conventional surface mix was placed as a control mix for performance comparison as part of this project.

81

The construction of the surface lift occurred on June 17–18, 2007. The paving contractor was Lojac Inc., and the HMA production plant was located north of Springfield. The mix was produced in a batch plant (3-ton capacity). Five passes of a steel roller was determined to be the appropriate rolling pattern from a test strip. Mixture was placed above 300oF. A 500-ft test section was identified 4,300 feet from the beginning of the project. 5.3.2 Mix Design

The TnDOT prepared the virgin mix design and 15% RAP mix design for the 4.75 mm HMA using the Marshall method with 75 blows. Note that TnDOT continues to use the Marshall method to design all asphalt mixtures. In April 2008, the TnDOT lab started with a mix of limestone screenings that produced a mixture with very high VMA (>20%). The final mix design for the virgin mix increased the natural sand to reduce the VMA. The final mix designs were completed in June 2007. A copy of each mix design is included in the Appendix. Table 5.5 summarizes the approved mix design for the virgin mix. Table 5.6 summarizes the approved mix design for the mix with RAP.

TABLE 5.5 Tennessee Validation Project 4.75 mm Virgin Mix Design Summary Mix Type Proposed AASHTO Criteria ACS-HM (surface mixture) Mix Size 4.75 mm NMAS 4.75 mm NMAS Binder Type PG 64-22 Binder Content 6.8 %

Aggregate Blend Nat.Sand=15% max. if FAA<45

75% screenings (#10-hard Aggr USA) 10% screenings (#10-soft Aggr USA) 15% natural-sand (Ingram Mtls)

Target Gradation 30%–55% passing 1.18 mm Sieve 6%–13% passing 0.075 mm Sieve

58% passing 1.18 mm Sieve 12.1% passing 0.075 mm Sieve

Aggregate Properties

FAA = 45 (min) SE = 40 (min) Not reported

Air Voids 4.0%–6.0% (Ndes=75 gyrations) 90.5 max (%Gmm @ Nini) Va=4.0% at 75-blow Marshall

Volumetric Properties

Vbe 12.0% –15.0% VMA 16.0 (note 1) VFA 65–78 (note 1) D:B ratio 1.0-2.0

Vbe=15.1 VMA=19.1 VFA = 79.0 D:B ratio=1.8

Moisture Susceptibility Not tested, not required based on asphalt

binder content Note 1 – current AASHTO criteria 5.3.3 Sampling and Testing Summary

The two 4.75 mm HMA mixtures were produced and placed June 17–18, 2008. NCAT staff was on the project site with a mobile laboratory to collect loose mix samples at the plant, compact production mixture to measure field lab volumetric properties, locate and cut cores, and measure

82

in-place friction. Samples were taken back to the NCAT laboratory for extracted material proportions, moisture susceptibility, rutting, and permeability testing. Table 5.7 and Table 5.8 summarize the production quality control test results performed by NCAT on each plant-produced mixture.

TABLE 5.6 Tennessee Validation Project 4.75 mm RAP Mix Design Summary Mix Type Proposed AASHTO Criteria ACS-HM (surface mixture with RAP) Mix Size 4.75 mm NMAS 4.75 mm NMAS Binder Type PG 64-22 Binder Content 6.8 %

Aggregate Blend Nat.Sand=15% max. if FAA<45

60% screenings (#10-hard Aggr USA) 10% screenings (#10-soft Aggr USA) 15% natural-sand (Ingram Mtls) 15% RAP (pass 5/16 Lojac)

Target Gradation 30%–55% passing 1.18 mm Sieve 6%–13% passing 0.075 mm Sieve

56% passing 1.18 mm Sieve 12.1% passing 0.075 mm Sieve

Aggregate Properties

FAA = 45 (min) SE = 40 (min) Not reported

Air Voids 4.0%–6.0% (Ndes=75 gyrations) 90.5 max (%Gmm @ Nini) Va=4.0% at 75-blow Marshall

Volumetric Properties

Vbe 12.0-15.0%% VMA 16.0 % min.(note 1) VFA 65-78 (note 1) D:B ratio 1.0-2.0

Vbe = 15.0 VMA = 19.0 VFA = 79% D:B ratio=1.8

Moisture Susceptibility Not tested, not required based on asphalt

binder content Note 1 – current AASHTO criteria

TABLE 5.7 NCAT Field Sampling and Testing for the Tennessee Validation Project

(Virgin Mix)

Test [no. of samples / no. of replicates] Mix Design Target Virgin Mix

Production QC (note-1)

Mixture Va–Lab(%Gmm@Ndes)[3/6] 4.0 4.6 – 5.9 Gmm [3/1] 2.389 2.398 – 2.407 Binder Content – by Ignition Method (Pb) [3/2] 6.8 7.5 – 7.7

Gradation – Washed from Ignition Samples [3/2] 58 pass 1.18 12.1 pass 0.075

50 – 51 11.7 – 13.4

Vbe VMA VFA D:B ratio

15.1 19.1 79.0 1.8

14.9 – 15.3 19.9 – 20.5 72.8 – 75.8 1.8 – 1.9

83

Moisture Susceptibility (TSR) [3/1] Not tested 0.68 – 0.75 Rut Testing – by APA (note-2) [3/2] 4.5 mm Lab Permeability from Field Cores (cm/sec) [8/1] 160 x 10-5 In-place Va – from Cores [8/1] 11.9avg, 7.5 – 14.2

Surface Friction – by DFT and CTM [8/3,2 (note-3)] DFT20 0.25 - 0.35 MPD 0.16 – 0.33 mm

Note-1 NCAT lab density results based on Ndes at 125 gyrations to match 4% Va. Note-2 Tested at design Va and at 7% voids. Note-3 Three replicates for DFT and two replicates for CTM.

TABLE 5.8 NCAT Field Sampling and Testing for the Tennessee Validation Project (15%

RAP Mix)

Test [no. of samples / no. of replicates] Mix Design Target 15% RAP

Production QC (note-1)

Mixture Va–Lab(%Gmm@Ndes) [3/6] 4.0 3.5 – 4.5 Gmm [3/1] 2.380 2.393 – 2.411 Binder Content – by Ignition Method (Pb) [3/2] 6.8 7.2 – 7.3

Gradation – Washed from Ignition Method [3/2] 56 pass 1.18 12.1 pass 0.075

52 – 54 13.2 – 14.1

Vbe VMAVFA D:B ratio

15.0 19.079.0

1.8

14.3 – 15.0 18.4 – 19.077.7 – 79.7

2.0 – 2.2 Moisture Susceptibility (TSR) [3/1] Not tested 0.67 – 0.79 Rut Testing – by APA (note-2) [3/2] 3.3 mm Lab Permeability from Field Cores (cm/sec) [8/1] 140 x 10-5 In-place Va – from Cores [8/1] (note 4) 11.7 avg, 10.7 – 12.7

Surface Friction – by DFT and CTM [8/3,2 (note-3)] DFT20 0.28 – 0.33 MPD 0.19 – 0.33 mm

Note-1 NCAT lab density results based on Ndes at 125 gyrations to match 4% Va. Note-2 Tested at design Va and at 7% voids. Note-3 Three replicates for DFT and two replicates for CTM. Note-4 One replicate measured Va=20.1 and was not included in the analysis. 5.3.4 Analysis of Mix Design and Production Test Results 5.3.4.1 Virgin Aggregate Mixture. The binder content determined by the 75-blow Marshall mix design procedure was 6.8%. The range of binder contents of the plant-produced virgin mixture measured by the ignition oven was 7.5% to 7.7%. The plant-produced binder contents were consistently 0.7% to 0.9% above the target established by the mix design.

The mix design gradation of the virgin mix design was finer than the recommended control points on the 1.18 mm sieve. However, the plant-produced virgin mix was coarser than the mix design target for the 1.18 mm control sieve with a tight range of 50% to 51% passing.

84

The amount of aggregate passing the 0.075 mm sieve was slightly above the mix design target, at an average of 12.7% with a good range of 11.7% to 13.4% passing.

For the TnDOT 4.75 mm virgin aggregate mixture, the mix design Va was 4.0% using a 75-blow Marshall design procedure. The NCAT lab attempted to match the volumetric properties of the mix design, but even at 125 gyrations with the SGC, the lab-compacted Va ranged from 4.6% to 5.9%. The Vbe for the plant-produced mix ranged from 14.9% to 15.3%, which was at the high end of the recommended range, but consistent with the design target of 15.0%. Likewise, VMA results and VFA were also very high. Although these volumetric properties may help this mixture resist reflection cracking from the underlying pavement, they raised concern about the potential for rutting of the mixture. However, APA test results were only 4.9 mm for specimens at 5.1% Va and 4.0 mm for specimens at 7.0% Va. These APA results would correspond to about 6 to 8 mm rutting in the MVT, which are very good compared to the results from the laboratory phase of this study.

The eight cores cut after field compaction had an average measured in-place Va of 11.9%, with a range of in-place voids from 7.5% to 14.2%. Six of the cores measured at or above 12.0%. Laboratory permeability measurements on cores were as high as 205 x 10-5 cm/sec. The range in permeability for the cores on this project was consistent with the range of in-place Va. Two cores with lower voids had lower permeability. The pavement may be more susceptible to permeability related distress in areas with more than 12.0% in-place Va. The measured lift thickness of the virgin mix cores varied greatly from 18.6 to 31.8 mm and met or exceeded the intended thickness of 19 mm. The variation in lift thickness implies that the existing pavement surface was irregular and may account for the wide range of measured in-place Va.

Moisture sensitivity of the plant-produced virgin mix yielded an average TSR of 0.71 with a range of 0.68 to 0.75 based on IDT tests of laboratory-prepared specimens at 7.0% to 7.3% Va. The TSR value was based on an average conditioned tensile strength of 157 psi and an average unconditioned tensile strength of 220 psi. No anti-strip treatment was required for this mixture based on TnDOT HMA criteria.

Friction characteristics of the virgin 4.75 mm HMA surface were measured with the DFT and CTM tests. The dynamic friction, based on the measured DFT20 values, ranged from 0.25 to 0.35 for the eight test locations. The MPD measured with the CTM ranged from 0.16 to 0.33 mm, which is in the typical range for fine-graded HMA with small NMAS aggregates. The DFT measurements reflect initial post-construction surface conditions. The high asphalt binder film on the surface creates lower friction results. Once the binder film is worn off by traffic, friction characteristics typically improve. 5.3.4.2 Mixture with 15% RAP. The binder content determined by the 75-blow Marshall mix design procedure was 6.8%. The Pbe was computed to 6.8%, so the combined virgin and RAP aggregates are not absorptive. The range of binder contents of the plant-produced mixture with 15% RAP measured by the ignition oven was 7.2% to 7.3%, consistently about 0.4% above the target established by the mix design.

85

The target gradation of the mix design with RAP was close to the recommended upper control points. The percent passing the 1.18 mm sieve was just above the upper control point, and the amount passing the 0.075 mm sieve was near the upper control point. Like the virgin Tennessee field mix, the gradation of the plant-produced mix with RAP was coarser than the mix design target for the 1.18 mm control sieve and finer than the mix design target for the 0.075 mm control sieve. The amount of aggregate passing the 1.18 mm sieve was 3% below the target, with a tight range of 52% to 54% passing. The amount of aggregate passing the 0.075 mm sieve was above the target at an average of 13.7%, with a range of 13.2% to 14.1% passing.

For the TnDOT 4.75 mm mixture with RAP, the mix design Va was 4.0% using a 75-blow Marshall design procedure. Field results tested by NCAT were based on 125 gyrations. With this compactive effort, the Va ranged from 3.5% to 4.5%, which was within a normal production specification tolerance of +/-1.0%. The Vbe of the plant-produced mix with RAP ranged from 14.3% to 15.0%. VMA ranged from 18.4% to 19.0%; VFA ranged from 78 to 80%. At face value, like the virgin TnDOT mix, these results for Vbe, VMA, and VFA are high and raise some concerns. It is not possible to precisely determine how these mix properties may have changed if the mix were compacted to 75 gyrations and designed for 6.0% Va, but a rough estimate is that VMA would have been in the range of 21% to 21.5%, asphalt content would decrease by about 0.9%, and VFA would probably be in the range of 71 to 72. The Vbe range would likely be high, in the range of 15.0% to 15.6%.

Rutting of the plant-produced 4.75 mm HMA with RAP from the APA test resulted in rut depths of 2.6 mm for specimens at 4.0% Va and 3.9 mm for specimens at 7.0% Va. Based on Figure 3.3, these APA results would correlate to about 3.0 to 5.5 mm, respectively, in the MVT. Therefore, despite the high Vbe results, the 4.75 mm mixture with RAP appears to be very rut resistant. This may be partly attributed to added stiffness from the RAP binder.

The eight cores cut after field compaction had an average in-place Va of 11.7% and a range from 10.7% to 20.1%. Seven of the cores measured between 10.7% and 12.7%. One core with 20.1% Va was not included in the computed average. Permeability of the seven cores taken from the test section was measured in the laboratory. Core permeability results ranged from 95 x 10-5 to 190 x 10-5 cm/sec, similar to the results for virgin TnDOT mix. The measured lift thickness of the cores with RAP varied greatly, from 19.2 to 31.2 mm and met or exceeded the intended thickness of 19 mm. There was no correlation between core thickness and core density.

Moisture sensitivity of the plant-produced mix with RAP yielded an average TSR of 0.72 with a range of 0.67 to 0.79 based on specimens with 6.7% to 7.2% Va. The average conditioned tensile strength was 173 psi, and the average unconditioned tensile strength was 239 psi. Friction characteristics of the 4.75 mm HMA with RAP surface were measured with the DFT and CTM tests. The dynamic friction based on the measured DFT20 values ranged from 0.28 to 0.33 for the eight test locations. The MPD measured with the CTM ranged from 0.19 to 0.33 mm. The DFT measurements reflect initial post-construction surface conditions. The asphalt binder film on the surface creates lower friction results. Once the binder film is worn off by traffic, the friction characteristics typically improve.

86

5.3.4.3 Virgin Mix – RAP Mix Comparison. Both mixtures were produced with 0.4% to 0.5% more asphalt binder than targeted in the mix design. The dust content in the mix with RAP (13.7%) was higher than the 12.7% for the virgin mix. Lab voids from plant-produced mix were lower for the mix with RAP (4.0%) compared to the virgin mix (5.1%). The higher dust content in the mix with RAP may account for the lower lab-compacted air void results. Overall, the in-place Va were slightly better for the mix with RAP, but the results for both mixtures were high and consequently had high permeability results, which raises concern. Both mixtures had poor moisture susceptibility results, but good rutting resistance results. 5.4 Minnesota Field Validation Project 5.4.1 Project Description The Minnesota DOT selected Section 6 of the MnRoad mainline experiment in the west-bound direction of I-94 to evaluate their 4.75 mm HMA mixture. Section 6 is 500 ft long and includes both the inside and outside lanes. The climate zone for this location is wet-freeze. The traffic level is monitored by the MnRoad research plan weigh-in-motion sensors, which typically log 600,000 ESALs in the driving lane annually. The research section would be designed for approximately 12 to 15 million ESALs for a 20-year design period. Therefore, this project is in the highest category for 4.75 mm mix designs (3.0 to 30 million ESAL). The posted speed for the section is 70 mph. The plans called for placing two 1-in lifts of 4.75 mm mixture over 5 inches of jointed-doweled PCC with a 15-ft joint spacing. The actual construction placed a single 2-inch lift of the 4.75 mm mixture. There was no specific control mix for comparison. The 4.75 mm surface is one of multiple sites along the research project.

The construction of the 4.75 mm test section occurred on October 30, 2008. The mix was produced in a drum plant and paved under tight experimental QC control. 5.4.2 Mix Design In June and July 2007, NCAT worked with MnDOT staff to prepare mix designs with taconite tailings and man-sand. The source of taconite tailings changed before the test sections were ready for construction the following year, so the mix designs were not used. The MnDOT prepared another mix design in July 2008. A copy of the mix design is included in the Appendix. Table 5.9 summarizes the approved mix design. 5.4.3 Sampling and Testing Summary The MnRoad 4.75 mm HMA was produced on one day of paving. NCAT staff was on the project site with a mobile laboratory to collect loose mix samples, compact production mixture to measure field-lab volumetric properties, obtain cores, and measure surface friction. The NCAT staff coordinated the 4.75 mm field sampling and testing requirements with the MnRoad experiment plan. Samples were taken back to the NCAT laboratory for extracted material

87

properties, moisture susceptibility, rutting, and permeability testing. Table 5.10 summarizes the production quality control test results performed by NCAT on the plant-produced mixture.

TABLE 5.9 Minnesota Validation Project 4.75 mm Mix Design Summary Mix Type Proposed AASHTO Criteria MnDOT SPWEB440F Special Mix Size 4.75 mm NMAS 4.75 mm NMAS Binder Type PG 64 -34 (polymer modified) Binder Content 7.4%, Pbe=6.9

Aggregate Blend 55% Taconite tailings (Mintac) 10% Taconite tailings (Ispat) 35% Man-sand (Loken)

Target Gradation 30%–55% passing 1.18 mm Sieve 6-13% passing 0.075 mm Sieve

51% passing 1.18 mm Sieve 7.7% passing 0.075 mm Sieve

Aggregate Properties FAA = 45 (min) SE = 45 (min) Nat.Sand=15(max) if FAA<45

FAA = 47 SE = 83 N/A

Air Voids 4.0%–6.0% (Ndes=75 gyrations) 89.0 max (%Gmm @ Nini)

Va=3.9% at Ndes =75 gyrations Not reported

Volumetric Properties Vbe 11.5-13.5 VMA 16.0 min. (note 1) VFA 65-78 (note 1) D:B ratio 1.5-2.0

Vbe=16.4 VMA=20.3 VFA 80.8 D:B ratio =1.1

Moisture Susceptibility TSR=0.82 @ Va = 9.0% Note 1 – current AASHTO criteria TABLE 5.10 NCAT Field Sampling and Testing for the Minnesota Validation Project Test [no. of samples / no. of replicates] Mix Design Target Production QC Mixture Va – Lab (%Gmm@Ndes) [3/2] 3.9 2.9 – 3.9 Gmm [3/1] 2.551 2.532 – 2.546 Binder Content –by ignition method (Pb) [3/note-1] 7.4% 8.8 – 9.1

Gradation–washed from ignition samples [3/note 1] 51 pass 1.18 7.7 pass 0.075

54 – 60 8.5 – 9.9

Vbe VMAVFA D:B ratio

16.4 20.380.8

1.1

17.9 – 18.5 21.0 – 22.182.7 – 85.4

1.1 – 1.3 Moisture Susceptibility (TSR) [3/1] 0.82 (9%) 0.68 – 0.82 Rut Testing – by APA [3/note-2] 5.3 mm Lab Permeability from field cores (cm/sec) [6/1] 5 x 10-5 In-place Va – from cores [6/1] 6.6 avg, 4.9 – 8.0

Surface Friction – by DFT and CTM (note-3) [10/3] DFT20 0.34 – 0.49 MPD 0.13 – 0.18 mm

Note-1 Four replicates for samples 1 & 2 and 2 replicates for sample 3

88

Note-2 Two replicates at design Va and 2 replicates at 7% Va Note-3 Five tests randomly spaced in each lane 5.4.4 Analysis of Mix Design and Production Test Results The binder content determined by the Ndes 75-gyration mix design procedure was 7.4%. The Pbe is 0.5% lower, indicating a moderate amount of absorption in the combined aggregate. The range of binder contents of the plant-produced mixture measured by the ignition oven was 8.8% to 9.1%, consistently more than 1.0% above the mix design target.

The target gradation of the mix design was within the proposed control points for the 1.18 mm and 0.075 mm sieves. The gradation of the plant-produced mix was finer than the mix design target for both control sieves. The amount of aggregate passing the 1.18 mm sieve during production was 5% above the target. The amount of aggregate passing the 0.075 mm sieve was 2% above the target.

For the MnDOT 4.75 mm mixture, the final mix design Va was 3.9%. The Va from plant-produced, lab-compacted mixture ranged from 2.9% to 3.9%, with an average of 3.6%. The increase in binder and dust contents are probable factors in the lower Va measured in the production mixture. The mix design Vbe was also 3% above the proposed range of 11.5% to 13.5%. During production, the Vbe results were much higher, ranging from 17.9% to 18.5%. The mixture was designed with a high VMA of 20.3%. The range of VMA from the plant-produced mixture was even higher—21.0 to 22.1%. The mix design VFA was 3% above the current AASHTO mix design specification range of 65 to 78. During production, the VFA further increased due to the higher asphalt binder content and VMA.

Rutting of the plant-produced 4.75 mm HMA tested with the APA test resulted in an average rut depth of 4.6 mm for specimens at 3.4% Va and 6.1 mm for specimens at 7.1% Va. Based on the correlation shown in Figure 3.3, these APA results would be approximately 6.5 mm and 9.5 mm, respectively, for the MVT. The Minnesota 4.75 mm mix rutting performance is better than expected for the high Vbe results. The good rutting resistance may be partly attributed to the use of the polymer modified PG 64-34 binder and the very angular, rough textured taconite aggregate.

Three cores were taken in each lane. The six cores had an average in-place Va of 6.6%, with a range of 4.9% to 8.0%. These in-place density results are very good for a field-compacted 4.75 mm mixture. As a result, the permeability results for the six cores were very low, less than 10 x 10-5 cm/sec. The total measured lift thickness for the one 2-inch lift ranged from 2.62 to 2.71 inches, well above the intended thickness.

Moisture sensitivity tests on the plant-produced mix yielded an average TSR of 0.76 with a range of 0.68 to 0.82 based on IDT tests of laboratory-prepared specimens at 7.1% to 7.3% Va. The average conditioned IDT strength was 78 psi, and the average unconditioned strength was 103 psi. The field TSR values are lower than the reported mix design TSR of 0.82 for 9% Va.

89

Friction characteristics of the 4.75 mm HMA surface were measured with the DFT and

CTM tests. The dynamic friction, based on the measured DFT20 values, ranged from 0.34 to 0.49 for the 10 test locations. The MPD measured with the CTM ranged from a 0.13 to 0.18 mm. The DFT measurements are better than the measurements at the Tennessee project and reflect the angular shape of the taconite aggregate tailings. The level of surface texture (MPD) is normal for fine-graded HMA with small NMAS aggregates. 5.5 Summary Analysis of Field Validation Results This section summarizes the mix designs and field results from the validation projects to determine how they complied with proposed mix design criteria. The summary analysis also determines if any deviations from the proposed mix design criteria are common across the projects and warrant a re-evaluation of the criteria. Table 5.11 summarizes the mix designs for the projects.

TABLE 5.11 Summary of Mix Designs for Validation Projects Mix Design Property Alabama Missouri Tennessee Minnesota

Mix Level 65 gyrations 50 gyrations 75-blow 75 gyrations Design Traffic (estimate) 1M ESAL 0.3M ESAL 1M ESAL 12M ESAL Binder Content (% of mix) 6.8 6.4 6.8 7.4 Effective Binder Content 6.5 5.4 6.8 6.9 1.18 mm Target (% passing) 47 48 58 51 0.075 mm Target (% passing) 6.0 7.6 12.1 7.7 Va (%) 3.3 4.0 4.0 3.9 VMA (%) 18.0 16.3 18.0 20.3 Vbe (%) 14.7 12.2 15.1 16.4 VFA (%) 81.8 75.2 79.0 80.8 D:B ratio 0.9 1.4 1.8 1.1 FT (microns) 9.3 7.1 6.5 8.5 TSR 0.85 Not measured Not measured 0.82 5.5.1 Compliance with Mix Design Standards Mixes for two of the four field projects were designed by the agencies to compaction levels different from the AASHTO mix design standards. Alabama’s use of 65 gyrations for Ndes fits within the range of the standards, but the mixture only measured 3.3% Va. The State of Tennessee used the Marshall 75-blow mix design standard. To match the Tennessee mix design, the NCAT laboratory applied 125 gyrations (Ndes) to achieve the target 4.0% Va.

Gradations – Figure 5.1 shows the average gradation during production for each project. All four mixtures were fine-graded mixtures. The Alabama mixture did not comply with the NMAS criteria. The mixture from Missouri was substantially finer than the other mixes between the 1.18 mm and the 0.075 mm sieves. All four validation mixes were designed and produced

90

near the upper control point on the 1.18 mm sieve. Each of the mix designs used a minimum of three aggregate stockpiles to blend into the combined gradation. The very fine gradations of these mixtures are a common, but not exclusive, characteristic for most 4.75 mm mixtures. The very fine gradations offer some advantages, such as the ability to place the mixtures in very thin applications with excellent smoothness, suitable for feathering, provide good joints, and maintain low permeability. During production of the mixtures, gradations were generally even finer than the mix designs. Three of the plant-produced mixes were an average of 5% finer than the mix design target on the 1.18 mm sieve. All four projects met the gradation criteria for aggregate passing the 0.075 mm sieve, but three of mixtures had significant increases in the amount of dust during production.

0

10

20

30

40

50

60

70

80

90

100

Sieve (mm)

Perc

ent P

assi

ng

ALMOTNMNcontrol ptsMDL

9.54.752.361.180.075 0.300

FIGURE 5.1 Average Plant-Production Gradation for Field Validation Projects

Table 5.12 summarizes the properties for the plant-produced mixtures. Results for each property are summarized by showing the average followed by the range. Table 5.13 identifies which measured properties from the field validation projects satisfied the mix design criteria.

Volumetrics – The limited number of samples obtained for field validation 4.75 mm mixtures indicate that, in general, the contractors maintained reasonable control of laboratory-compacted Va during production. It is evident that, except for the Alabama mix, the asphalt contents had to be increased substantially over the mix design targets, even with the higher production dust contents.

The laboratory phase of this study recommended the use of criteria for Vbe in place of VMA and VFA. The Vbe of the Missouri mix was designed and produced within the recommended Vbe range. Three other three mixes were produced at or above the maximum recommended mix design criteria for Vbe for their respective traffic categories, either by design or by additional asphalt during production. The Minnesota mix exceeded the proposed limit of

91

13.5% by more than 3%. However, despite the high Vbe results, the rutting tests on the validation mixtures were acceptable.

Although VMA and VFA criteria are not recommended for continued use for 4.75 mm mixtures, these properties were analyzed to maintain continuity with the historical criteria. As with most of the Phase I mix designs, the verification mixes easily met the current AASHTO minimum VMA criteria of 16.0%. During production, three of the mixtures had an increase in VMA due to gradation shifts away from the maximum density line, even with higher dust contents. In contrast, three mix designs exceeded the current AASHTO recommended upper limits for VFA. The Missouri mix was within the VFA range allowed for low-traffic projects for the mix design and during production. The VFA results are consistent with the Vbe results noted above.

TABLE 5.12 Summary of Plant-Produced Mixes for Validation Projects Field Property

(average, range) Alabama Missouri Tennessee

(virgin mix) Minnesota

Mix Level 65 gyrations 50 gyrations 125 gyrations (note1)

75 gyrations

Binder Content (% of mix) 6.9, 6.7-7.1 7.2, 6.8-7.4 7.6, 7.5-7.7 9.0, 8.8-9.1

1.18 mm Sieve (% passing) 52, 51-55 54, 48-58 50, 50-51 56, 54-60

0.075 mm Sieve (% passing) 9.2, 8.3-11 12.1, 11.4-12.5 12.7, 11.7-13.4 9.0, 8.7-9.9

Va (lab compacted) 2.8, 1.9-3.7 4.2, 4.1-4.5 5.1, 4.6-5.9 3.6, 2.9-3.9

VMA (%) 18.3, 17.8-18.7 17.2, 16.6-17.7 20.3, 19.9-20.5 21.6, 21.0-22.1

Vbe (%) 15.6, 14.4-16.2 13.0, 12.5-13.3 15.1, 14.9-15.3 18.2, 17.9-18.5

VFA (%) 85.1, 80.6-88.1 75.3, 74.3-76.5 74.5, 72.8-75.8 84.2, 82.7-85.4

In-place Va (%) 11.7, 9.5-13.2 10.1, 9.2-11.9 11.9, 7.5-14.2 6.6, 4.9-8.0

D:B ratio 1.4, 1.2-1.8 2.1, 2.0-2.2 1.9, 1.8-1.9 1.2, 1.1-1.3

FT (microns) 7.5 5.0 6.7 8.6

TSR 0.80 0.70, 0.66-0.74 0.71, 0.68-0.75 0.76, 0.68-0.82

APA (mm) 13 6.7 (Va=4%) 4.0 (Va=7%) 6.1 (Va=7%)

Permeability (cm/sec) 90 x 10-5 40x10-5 140x10-5 <10x10-5

DFT20 Not measured Not measured 0.25-0.35 0.34-0.49

CTM (mm) Not measured 0.17-0.22 0.16-0.33 0.13-0.18 Note-1 Field mix compacted to 125 gyrations required to match 75-blow Marshall mix design at 4.0% Va.

The D:B ratio of the mix designs reflected the full range of the criteria. For comparison, the computed FT for the mix designs ranged from a somewhat low value of 6.5 microns to a more reasonable value of 9.3 microns. During production, the computed D:B ratio for the Missouri mix increased above 2.0 and the FT dropped to 5.0 microns. This raises some concern about the mixture’s durability. For projects in the traffic category of 3.0 to 30 million ESALs, the authors recommended a minimum D:B ratio of 1.5. This recommendation was based on the trend observed in the lab phase that 4.75 mm mixtures with higher D:B ratios had much better

92

rutting resistance. However, the Minnesota validation mixture proved that mixtures with low D:B ratios could have good rutting resistance if other mix design criteria are met. Requiring a minimum D:B ratio of 1.5 for this traffic level could be too restrictive on mix designs and could also reduce mixture durability. Therefore, the D:B ratio should be 1.0 to 2.0 for all traffic levels.

TABLE 5.13 Mix Design Criteria Validation Summary

Mix criteria

Traff

ic Le

vel

(ESA

Ls)

Curre

nt AA

SHTO

Preli

m.

Reco

mm.

Crite

ria

Alab

ama

Miss

ouri

Tenn

esse

e

Minn

esota

Ndes Gyrations <0.3M <3.0M >3.0M

50 75 75

No change 65 OK 50 OK 125 (note 1) <3.0M HIGH 75-blow non-standard

75 OK

FAA (note-3)

<0.3M <3.0M >3.0M

40 45 45

Design OK

Design OK

Design not reported

Design OK

Natural Sand (if FAA<45)

<0.3M <3.0M >3.0M

15 max 15 max

Design n/a

Design n/a

Design OK

Design n/a

SE (note-3)

<0.3M <3.0M >3.0M

40 40 45

No change Design Not reported

Design Not reported

Design Not reported

Design OK

Gradation Control 1.18 mm

<0.3M <3.0M >3.0M

30-60 30–55 Design OK Plant OK

Design OK Plant OK

Design HIGH Plant OK

Design OK Plant HIGH

Gradation Control 0.075 mm

6–12 6–13 Design OK Plant OK (note 2)

Design OK Plant OK (note 2)

Design OK Plant OK (note 2)

Design OK Plant OK (note 2)

Va (%Gmm@Ndes)

4.0 4.0–6.0 Design LO Plant LOW

Design OK Plant OK

Design OK Plant OK

Design OK Plant LOW

VMA <0.3M <3.0M >3.0M

16 min 16 min 16 min

Design OK Plant HIGH

Design OK Plant OK

Design OK Plant HIGH

Design HI Plant HIGH

Vbe <0.3M <3.0M >3.0M

12.0-15.0 12.0–15.0 11.5–13.5

Design OK Plant HIGH

Design OK Plant OK

Design OK Plant OK

Design HI Plant HIGH

%Gmm @ Nini (note-3)

<0.3M <3.0M >3.0M

91.5 max 90.5 max 89.0 max

No change Design OK

Design Not reported

Design n/a

Design Not reported

D:B Ratio <0.3M <3.0M

0.9–2.0 0.9–2.0

1.0–2.0 1.0–2.0

Design LO Plant OK

Design OK Plant HIGH

Design OK Plant OK

Design OK Plant HIGH

93

>3.0M 0.9–2.0 1.5–2.0 Note-1 Applied 125 gyrations to match 75-blow Marshall mix design volumetrics. Note-2 All production mixture increased the mineral fines over the mix design. Note-3 These mix design criteria were not examined as part of field validation.

Figure 5.2 summarizes the results of the moisture susceptibility testing on the plant-produced mixtures. The field mixes had TSR results in the range of 0.7 to 0.8, which would be considered marginal at best by most agency specifications. Missouri and Tennessee did not require moisture damage testing on the mix designs. TSR results can mask significant differences between the mixtures, as presented by the average IDT test results. For example, although the Alabama and Minnesota TSR values were higher, their average tensile strengths were the lowest. Conversely, the TSR of the Missouri mix was lower than for the Alabama and Minnesota mixes, but its average conditioned tensile strength was higher. Another issue with moisture damage testing of these fine-graded 4.75 mm mixtures is the possibility that the vacuum saturation process may cause damage within the specimens due to the rapid expansion of air within the small voids of the specimen. Therefore, TSR results of the validation mixes are inconclusive.

0

50

100

150

200

250

AL 65 gyr MO 50 gyr TN 75-blow MN 75 gyr

Validation Project

IDT

(psi

)

0.5

0.6

0.7

0.8

0.9

1

TSR

dry IDT wet IDT TSR

FIGURE 5.2 Average AASHTO T 283 Results for Plant-Produced Mixtures

In-place Air Void Contents – Since 4.75 mm mixtures are typically placed in lifts less than 1-in. thick, field in-place densities or Va are not usually measured. These mixes are often placed on irregular surfaces, so consistent thicknesses are usually not possible. Generally, agencies simply require inspectors to establish a set rolling pattern for the contractor to follow. Therefore, it was expected that in-place Va would be higher than the normal range of 6% to 8% for most HMA. The results of the density tests on the cores showed that several of the 4.75 mm mixtures had Va as high as 13% and 14%. The MnRoad project had much lower Va, which can likely be attributed to the extra care taken to build the short test section at MnRoad and the use of

94

a very thick lift (over 2.6 inches). Permeability tests on the cores from the projects showed that even at the high Va, the pavements were relatively impermeable.

Figure 5.3 shows a good relationship between the in-place Va and laboratory permeability for the cores from all the projects. This graph shows that to maintain an impermeable surface mix of less than 125 x 10-5 cm/sec, the field in-place Va should be kept below 12%.

0

50

100

150

200

250

300

0 2 4 6 8 10 12 14 16

In-Place Air Voids (%)

Perm

eabi

lity

(10

-5 c

m/s

ec)

FIGURE 5.3 In-Place Air Voids Versus Permeability for 4.75 mm Mixtures

Friction – Initial CTM values reflect the very fine surface texture created by the fine-

graded 4.75 mm mixtures. Hard polish resistant aggregate will be the key to retaining friction properties because the surface texture will not be a major contributing factor. The initial DFT friction values do not typically represent the maximum friction resistance of an HMA surface. The initial measured DFT friction values are negatively influenced by the thin film of asphalt binder on the surface of the aggregate particles. As the traffic wears the asphalt binder film off the surface, exposing more of the aggregate, the friction values typically improve. Once the maximum surface friction is reached after a few weeks or months of traffic, then the friction value is dependent on the polish resistance of the aggregate.

95

CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS 6.0 CONCLUSIONS AND RECOMMENDATIONS 6.1 Conclusions The objective of this research was to refine the current AASHTO criteria for 4.75 mm NMAS Superpave designed mixtures. In Phase I, 29 4.75 mm NMAS Superpave mix designs were prepared in the laboratory with materials from nine states. Each mix design was tested for permanent deformation, permeability, moisture damage susceptibility, and durability. Plant-produced mixtures from four other states were evaluated and served as baselines for performance. After the Phase I laboratory study, four of the original nine participating state highway agencies constructed projects to validate the proposed mix design criteria and establish field construction criteria. Based on the results of this research, the following conclusions were made with regard to the design of 4.75 mm mixtures:

• The design of 4.75 mm NMAS mixtures is largely dependent on the characteristics of available fine aggregates. In general, 4.75 mm mix designs should utilize at least three aggregate stockpiles to develop a suitable blend that can be adequately controlled during plant production. Most blends of available materials tend to result in very fine gradations, which generally have excessive voids in mineral aggregate (VMA). Consequently, 4.75 mm mixtures often have very high asphalt contents (>6.5% to 9.0%) and tend to be susceptible to permanent deformation.

• VMA and design asphalt content of 4.75 mm mixtures can be reduced by using coarser gradations (closer to the maximum density line) and increasing the dust content.

• The compactive effort used for the mix design should be consistent with the design traffic level. Fifty gyrations are suitable for low-traffic applications where rutting is not a concern either because traffic speeds are low or because the pavement will not carry heavy vehicles that may cause permanent deformation. Seventy-five gyrations should be suitable for most other applications of 4.75 mm mixtures.

• Using a design air void range of 4.0% to 6.0% has little effect on the VMA but will allow mix designers to reduce the asphalt content for a given aggregate blend when the VMA is well above 16.0%. This will improve resistance of 4.75 mm mixtures to permanent deformation. Mixtures with less than 13.5% volume of effective binder (Vbe) had better rutting resistance than mixtures with more than 13.5% Vbe.

• Additionally, resistance of 4.75 mm mixtures to permanent deformation can be improved by increasing the dust content, using aggregates with high angularity, and using stiffer binders. Mixtures with dust-to-binder ratios above 1.5 had lower average rutting rates than mixtures with less than a 1.5 dust-to-binder ratio.

• Susceptibility to moisture damage generally increased slightly with decreasing effective asphalt contents. Natural sand contents of over 15% appear to adversely affect moisture susceptibility, rutting susceptibility, and permeability.

96

• Laboratory permeability testing of the lab and field mixes demonstrated that fine-graded 4.75 mm NMAS mixtures are practically impermeable even at relatively high in-place air voids. Low permeability should help reduce exposure to moisture for these mixtures.

• Higher asphalt contents tended to increase fracture energy ratio. Based on the plots of fracture energy versus film thickness and dust-to-binder ratio, it is concluded that a 4.75 mm NMAS mixture’s ability to sustain resistance to cracking is a function of both asphalt content and dust content. Therefore, to assure good durability, 4.75 mm mix criteria should include a minimum Vbe and a maximum dust-to-binder ratio.

6.2 Recommendations Based on the results of this study, the following recommendations are provided:

• The current gradation limits on the 1.18 mm and 0.075 mm sieves should be adjusted. Limits placed on percent passing the 1.18 sieve should be 30% to 55%. Limits placed on percent passing the 0.075 mm sieve should be 6.0% to 13.0%.

• For mixes designed for over 0.3 million ESALs, the aggregate blend should contain no more than 15% natural sand and have a minimum fine aggregate angularity of 45 for improved rut resistance, moisture damage resistance, and to maintain low permeability.

• The target air void content for selecting the design binder content should be changed to a range from 4.0% to 6.0%. This will allow for a reduction in the design asphalt content for many 4.75 mm mixtures that have very high VMAs.

• Criteria for VMA and VFA should be replaced with minimum and maximum Vbe requirements. This is a more sensible approach when a range of design air voids is used. For less than 3.0 million design ESALs, a Vbe range of 12.0% to 15.0% is recommended. For 4.75 mm mixtures designed for projects over 3.0 million ESALs, a minimum Vbe of 11.5% and a maximum Vbe of 13.5% is recommended. These limits were based on fracture energy testing and rut testing for the minimum and maximum Vbe, respectively.

• The maximum %Gmm @ Nini requirement appears appropriate for both 4.0% and 6.0% design air voids. At this time it is recommended that current Gmm @ Nini criteria be maintained.

• The minimum dust-to-binder ratio should be increased slightly from 0.9 to 1.0. The maximum dust-to-binder ratio should be maintained at 2.0.

• No evidence was found that suggested adjusting the current sand equivalent minimum. At this time, the minimum sand equivalent criteria should be maintained.

• The moisture sensitivity test should be reviewed to reduce vacuum induced damage for 4.75 mm mixtures to account for the lower mixture permeability.

A summary of proposed mix design criteria is given in Table 6.1.

97

TABLE 6.1 Proposed Design Criteria for 4.75 mm NMAS Superpave Design Mixtures Design ESAL Range

(Millions) Ndes Minimum

FAA Minimum

SE Minimum

Vbe Maximum

Vbe

%Gmm@Nini D:B

Ratio

<0.3 50 40 40 12.0 15.0 ≤91.5 1.0 to 2.0 0.3 to ≤ 3.0 75 45 40 11.5 13.5 ≤90.5 1.0 to 2.0 3.0 to ≤ 30 100 45 45 11.5 13.5 ≤89.0 1.0 to 2.0

Gradation Limits

Sieve Size Max. Min. Design Va Range = 4.0% to 6.0% 12.5 mm --- 100 9.5 mm 100 95

4.75 mm 100 90 1.18 mm 30 55 0.075 mm 13 6

6.3 Recommended Applications, Advantages, and Disadvantages 4.75 mm mixtures and similar fine-graded mixtures have been used by many transportation agencies for a variety of pavement applications. The most common use of 4.75 mm mixtures is as a surface course on low traffic volume applications such as neighborhood streets. However, these mixes can also be effective as a thin maintenance overlay (less than one inch) on a variety of roadways. Other applications of 4.75 mm mixtures include parking lots, patching mixtures, and use as a leveling course to restore pavement cross slope and profile.

Advantages of 4.75 mm mixtures include the following:

• Thin lifts (3/4 to one inch, typically) • Excellent smoothness • Practically impermeable • Good use of fine aggregate materials, which are in surplus in many areas • Good use of fractionated fine RAP, which helps improve the stability of 4.75 mm

mixtures • Can be feathered and used in wedge applications • Good workability with hand tools

Disadvantages of 4.75 mm mixtures include the following:

• High asphalt contents. This can be offset some by including fractionated fine RAP in the

mix design. Even at high asphalt contents, these mixtures are very economical because they can cover a large area per ton.

• Low frictional resistance due to the low surface texture of the mixtures. Although using tough and high angularity fine aggregates can provide good skid resistance in dry conditions and in wet weather at slow speeds, 4.75 mm mixtures should not be used on heavy traffic, high speed roadways.

98

• Greater potential for permanent deformation. The solution for good durability in thin applications is a higher asphalt content, which makes the 4.75 mm mixtures more susceptible to rutting. Resistance to permanent deformation can be improved by 1) designing mixes with lower VMA, 2) using highly angular aggregates, RAP, and a stiffer polymer modified binder, and 3) by achieving good density/low in-place air voids during construction.

99

REFERENCES

1) Cooley, L. A., R. S. James, and S. M. Buchanan. Development of Mix Design Criteria for 4.75 mm Superpave Mixes. National Center for Asphalt Technology, Report No. 2002-4, February 2004.

2) Cooley, L. A., M. H. Hunter, and E. R. Brown. Use of Screenings to Produce HMA Mixtures. National Center for Asphalt Technology, Report No. 2002-10, October 2002.

3) Prowell, B. D. and J. E. Haddock. Superpave for Low Volume Roads and Base Mixtures. Journal of the Association of Asphalt Paving Technologists, Vol. 71, 2002.

4) Mogawer, W. S. and R. Mallick. Design of Superpave HMA for Low Volume Roads. The New England Transportation Consortium, December 2004. www.netc.uconn.edu/reports_listing.html#materials. Accessed December 6, 2005.

5) Habib, A., M. Hossain, R. Kaldate, and G. A. Flager. Comparison of Superpave and Marshall Mixtures for Low-Volume Roads and Shoulders. Transportation Research Board, Transportation Research Record 1609, 1998.

6) Engle, E. J. Superpave Mix Designs for Low-Volume Roads. Iowa Highway Research Board, October 2004. www.operationsresearch.dot.state.ia.us/reports/ihrb_by_number/tr400plus.html. Accessed December 6, 2005.

7) Experiences with Superpave on County Roads. Asphalt. Magazine of the Asphalt Institute, Vol. 19, No. 1, Spring 2004, pp. 22–24.

8) Asphalt in Pavement Maintenance. The Asphalt Institute. Manual Series No.16 (MS-16), March 1983.

9) Personal Communication with Donald Watson, National Center for Asphalt Technology, March 23, 2006.

10) Hasen, K. Pavement Preservation with Thin Overlays. Better Roads, Vol. 73, No. 6, June 2003, pp. 48–50.

11) Brown, E. R., M. R. Hainin, L. A. Cooley, and G. Hurley. NCHRP Report 531, Relationship of Air Voids, Lift Thickness, and Permeability in Hot Mix Asphalt Pavements. Transportation Research Board, National Research Council, Washington DC, 2004.

12) Moore, J. R. and B. D. Prowell. Evaluation of the Mix Verification Tester for Determining the Rutting Susceptibility of Hot Mix Asphalt. National Center for Asphalt Technology, Internal Report, 2006.

13) Mallick, R. B., L. A. Cooley, M. R. Teto, R. L. Bradbury, and D. Peabody. An Evaluation of Factors Affecting Permeability of Superpave Designed Pavements. National Center for Asphalt Technology, Report 03-02, June 2003.

14) Kim, Y. R. and H. Wen. Fracture Energy from Indirect Tension Testing. Journal of the Association of Asphalt Paving Technologists, Vol. 71, 2002.

15) Birgisson, B., C. Soranakom, J. A. L. Napier, and R. Roque. Simulation of Fracture Initiation in Hot-Mix Asphalt Mixtures. Transportation Research Board, Transportation Research Record 1849, 2003.

16) Prowell, D. P., J. Zhang, and E. R. Brown. NCHRP Report 539, Aggregate Properties and Performance of Superpave Designed Hot Mix Asphalt. Transportation Research Board. National Research Council, Washington DC, 2005.

100

UNCITED REFERENCES

1) Level One Mix Design: Materials Selection, Compaction, and Conditioning. SHRP-A-

408, Strategic Highway Research Program, National Research Council, Washington, DC, August 1994.

2) Kandhal, P. S. and L. A. Cooley. NCHRP Report 464, The Restricted Zone in the Superpave Aggregate Gradation Specification. Transportation Research Board. National Research Council, Washington DC, 2001.

3) Roberts, F. L., P .S. Kandhal, E. R. Brown, D. Y. Lee, and T. W. Kennedy. Hot Mix Asphalt Materials, Mixture Design and Construction. National Asphalt Pavement Association Rese arch and Education Foundation, 1996.

4) Stroup-Gardiner, M., D. Newcomb, W. Kussman, and R. Olsen. Characteristics of Typical Minnesota Aggregates, Transportation Research Record 1583, Transportation Research Board, National Research Council, Washington DC, 1997.

5) Kandhal, P. S., C. Lynn, and F. Parker. Test for Plastic Fines in Aggregates Related to Stripping in Asphalt Paving Mixtures, National Center for Asphalt Technology, Report No. 98-3, 1998.

6) Kandhal, P. S. and L. A. Cooley. NCHRP Report 508, Accelerated Laboratory Rutting Tests: Evaluation of the Asphalt Pavement Analyzer. Transportation Research Board. National Research Council, Washington DC, 2003.

7) Cooley Jr., L. A., B. D. Prowell, and E. R. Brown. Issues Pertaining to the Permeability Characteristics of Coarse-Graded Superpave Mixes. National Center for Asphalt Technology, Report 02-06, July 2002.

8) Roque, R., B. Birgisson, Z. Zhang, B. Sangpentngam, and T. Grant. Implementation of SHRP Indirect Tension Tester to Mitigate Cracking in Asphalt Pavements and Overlays. University of Florida, May 2002. www.dot.state.fl.us/research-center/Completed_StateMaterials.htm. Accessed February 2, 2006.

9) Herrin, M. Bituminous Patching Mixtures. Transportation Research Board, NCHRP Synthesis 64, 1979.

101

APPENDIX

Mix Designs for the Field Validation Projects

Alabama............................................................................... p. 102 Missouri............................................................................... p. 103 Tennessee............................................................................ p. 104 Minnesota............................................................................ p. 106

102

103

104

105

106